CINXE.COM

<!doctype html> <html lang="en-US" id="html"> <head> <meta charset="utf-8"> <meta http-equiv="x-ua-compatible" content="ie=edge"><script type="text/javascript">(window.NREUM||(NREUM={})).init={ajax:{deny_list:["bam.nr-data.net"]}};(window.NREUM||(NREUM={})).loader_config={licenseKey:"NRJS-2b0defbeb07862260b2",applicationID:"745178505"};;/*! For license information please see nr-loader-rum-1.283.2.min.js.LICENSE.txt */ (()=>{var e,t,r={122:(e,t,r)=>{"use strict";r.d(t,{a:()=>i});var n=r(944);function i(e,t){try{if(!e||"object"!=typeof e)return(0,n.R)(3);if(!t||"object"!=typeof t)return(0,n.R)(4);const r=Object.create(Object.getPrototypeOf(t),Object.getOwnPropertyDescriptors(t)),o=0===Object.keys(r).length?e:r;for(let a in o)if(void 0!==e[a])try{if(null===e[a]){r[a]=null;continue}Array.isArray(e[a])&&Array.isArray(t[a])?r[a]=Array.from(new Set([...e[a],...t[a]])):"object"==typeof e[a]&&"object"==typeof t[a]?r[a]=i(e[a],t[a]):r[a]=e[a]}catch(e){(0,n.R)(1,e)}return r}catch(e){(0,n.R)(2,e)}}},555:(e,t,r)=>{"use strict";r.d(t,{Vp:()=>c,fn:()=>s,x1:()=>u});var n=r(384),i=r(122);const o={beacon:n.NT.beacon,errorBeacon:n.NT.errorBeacon,licenseKey:void 0,applicationID:void 0,sa:void 0,queueTime:void 0,applicationTime:void 0,ttGuid:void 0,user:void 0,account:void 0,product:void 0,extra:void 0,jsAttributes:{},userAttributes:void 0,atts:void 0,transactionName:void 0,tNamePlain:void 0},a={};function s(e){try{const t=c(e);return!!t.licenseKey&&!!t.errorBeacon&&!!t.applicationID}catch(e){return!1}}function c(e){if(!e)throw new Error("All info objects require an agent identifier!");if(!a[e])throw new Error("Info for ".concat(e," was never set"));return a[e]}function u(e,t){if(!e)throw new Error("All info objects require an agent identifier!");a[e]=(0,i.a)(t,o);const r=(0,n.nY)(e);r&&(r.info=a[e])}},217:(e,t,r)=>{"use strict";r.d(t,{D0:()=>m,gD:()=>v,xN:()=>h});r(860).K7.genericEvents;const n="experimental.marks",i="experimental.measures",o="experimental.resources",a=e=>{if(!e||"string"!=typeof e)return!1;try{document.createDocumentFragment().querySelector(e)}catch{return!1}return!0};var s=r(614),c=r(944),u=r(384),l=r(122);const d="[data-nr-mask]",f=()=>{const e={feature_flags:[],experimental:{marks:!1,measures:!1,resources:!1},mask_selector:"*",block_selector:"[data-nr-block]",mask_input_options:{color:!1,date:!1,"datetime-local":!1,email:!1,month:!1,number:!1,range:!1,search:!1,tel:!1,text:!1,time:!1,url:!1,week:!1,textarea:!1,select:!1,password:!0}};return{ajax:{deny_list:void 0,block_internal:!0,enabled:!0,autoStart:!0},distributed_tracing:{enabled:void 0,exclude_newrelic_header:void 0,cors_use_newrelic_header:void 0,cors_use_tracecontext_headers:void 0,allowed_origins:void 0},get feature_flags(){return e.feature_flags},set feature_flags(t){e.feature_flags=t},generic_events:{enabled:!0,autoStart:!0},harvest:{interval:30},jserrors:{enabled:!0,autoStart:!0},logging:{enabled:!0,autoStart:!0},metrics:{enabled:!0,autoStart:!0},obfuscate:void 0,page_action:{enabled:!0},page_view_event:{enabled:!0,autoStart:!0},page_view_timing:{enabled:!0,autoStart:!0},performance:{get capture_marks(){return e.feature_flags.includes(n)||e.experimental.marks},set capture_marks(t){e.experimental.marks=t},get capture_measures(){return e.feature_flags.includes(i)||e.experimental.measures},set capture_measures(t){e.experimental.measures=t},capture_detail:!0,resources:{get enabled(){return e.feature_flags.includes(o)||e.experimental.resources},set enabled(t){e.experimental.resources=t},asset_types:[],first_party_domains:[],ignore_newrelic:!0}},privacy:{cookies_enabled:!0},proxy:{assets:void 0,beacon:void 0},session:{expiresMs:s.wk,inactiveMs:s.BB},session_replay:{autoStart:!0,enabled:!1,preload:!1,sampling_rate:10,error_sampling_rate:100,collect_fonts:!1,inline_images:!1,fix_stylesheets:!0,mask_all_inputs:!0,get mask_text_selector(){return e.mask_selector},set mask_text_selector(t){a(t)?e.mask_selector="".concat(t,",").concat(d):""===t||null===t?e.mask_selector=d:(0,c.R)(5,t)},get block_class(){return"nr-block"},get ignore_class(){return"nr-ignore"},get mask_text_class(){return"nr-mask"},get block_selector(){return e.block_selector},set block_selector(t){a(t)?e.block_selector+=",".concat(t):""!==t&&(0,c.R)(6,t)},get mask_input_options(){return e.mask_input_options},set mask_input_options(t){t&&"object"==typeof t?e.mask_input_options={...t,password:!0}:(0,c.R)(7,t)}},session_trace:{enabled:!0,autoStart:!0},soft_navigations:{enabled:!0,autoStart:!0},spa:{enabled:!0,autoStart:!0},ssl:void 0,user_actions:{enabled:!0,elementAttributes:["id","className","tagName","type"]}}},g={},p="All configuration objects require an agent identifier!";function m(e){if(!e)throw new Error(p);if(!g[e])throw new Error("Configuration for ".concat(e," was never set"));return g[e]}function h(e,t){if(!e)throw new Error(p);g[e]=(0,l.a)(t,f());const r=(0,u.nY)(e);r&&(r.init=g[e])}function v(e,t){if(!e)throw new Error(p);var r=m(e);if(r){for(var n=t.split("."),i=0;i<n.length-1;i++)if("object"!=typeof(r=r[n[i]]))return;r=r[n[n.length-1]]}return r}},371:(e,t,r)=>{"use strict";r.d(t,{V:()=>f,f:()=>d});var n=r(122),i=r(384),o=r(154),a=r(324);let s=0;const c={buildEnv:a.F3,distMethod:a.Xs,version:a.xv,originTime:o.WN},u={customTransaction:void 0,disabled:!1,isolatedBacklog:!1,loaderType:void 0,maxBytes:3e4,onerror:void 0,ptid:void 0,releaseIds:{},appMetadata:{},session:void 0,denyList:void 0,timeKeeper:void 0,obfuscator:void 0,harvester:void 0},l={};function d(e){if(!e)throw new Error("All runtime objects require an agent identifier!");if(!l[e])throw new Error("Runtime for ".concat(e," was never set"));return l[e]}function f(e,t){if(!e)throw new Error("All runtime objects require an agent identifier!");l[e]={...(0,n.a)(t,u),...c},Object.hasOwnProperty.call(l[e],"harvestCount")||Object.defineProperty(l[e],"harvestCount",{get:()=>++s});const r=(0,i.nY)(e);r&&(r.runtime=l[e])}},324:(e,t,r)=>{"use strict";r.d(t,{F3:()=>i,Xs:()=>o,xv:()=>n});const n="1.283.2",i="PROD",o="CDN"},154:(e,t,r)=>{"use strict";r.d(t,{OF:()=>c,RI:()=>i,WN:()=>l,bv:()=>o,gm:()=>a,mw:()=>s,sb:()=>u});var n=r(863);const i="undefined"!=typeof window&&!!window.document,o="undefined"!=typeof WorkerGlobalScope&&("undefined"!=typeof self&&self instanceof WorkerGlobalScope&&self.navigator instanceof WorkerNavigator||"undefined"!=typeof globalThis&&globalThis instanceof WorkerGlobalScope&&globalThis.navigator instanceof WorkerNavigator),a=i?window:"undefined"!=typeof WorkerGlobalScope&&("undefined"!=typeof self&&self instanceof WorkerGlobalScope&&self||"undefined"!=typeof globalThis&&globalThis instanceof WorkerGlobalScope&&globalThis),s=Boolean("hidden"===a?.document?.visibilityState),c=/iPad|iPhone|iPod/.test(a.navigator?.userAgent),u=c&&"undefined"==typeof SharedWorker,l=((()=>{const e=a.navigator?.userAgent?.match(/Firefox[/\s](\d+\.\d+)/);Array.isArray(e)&&e.length>=2&&e[1]})(),Date.now()-(0,n.t)())},687:(e,t,r)=>{"use strict";r.d(t,{Ak:()=>c,Ze:()=>d,x3:()=>u});var n=r(836),i=r(606),o=r(860),a=r(646);const s={};function c(e,t){const r={staged:!1,priority:o.P3[t]||0};l(e),s[e].get(t)||s[e].set(t,r)}function u(e,t){e&&s[e]&&(s[e].get(t)&&s[e].delete(t),g(e,t,!1),s[e].size&&f(e))}function l(e){if(!e)throw new Error("agentIdentifier required");s[e]||(s[e]=new Map)}function d(e="",t="feature",r=!1){if(l(e),!e||!s[e].get(t)||r)return g(e,t);s[e].get(t).staged=!0,f(e)}function f(e){const t=Array.from(s[e]);t.every((([e,t])=>t.staged))&&(t.sort(((e,t)=>e[1].priority-t[1].priority)),t.forEach((([t])=>{s[e].delete(t),g(e,t)})))}function g(e,t,r=!0){const o=e?n.ee.get(e):n.ee,s=i.i.handlers;if(!o.aborted&&o.backlog&&s){if(r){const e=o.backlog[t],r=s[t];if(r){for(let t=0;e&&t<e.length;++t)p(e[t],r);Object.entries(r).forEach((([e,t])=>{Object.values(t||{}).forEach((t=>{t[0]?.on&&t[0]?.context()instanceof a.y&&t[0].on(e,t[1])}))}))}}o.isolatedBacklog||delete s[t],o.backlog[t]=null,o.emit("drain-"+t,[])}}function p(e,t){var r=e[1];Object.values(t[r]||{}).forEach((t=>{var r=e[0];if(t[0]===r){var n=t[1],i=e[3],o=e[2];n.apply(i,o)}}))}},836:(e,t,r)=>{"use strict";r.d(t,{P:()=>c,ee:()=>u});var n=r(384),i=r(990),o=r(371),a=r(646),s=r(607);const c="nr@context:".concat(s.W),u=function e(t,r){var n={},s={},l={},d=!1;try{d=16===r.length&&(0,o.f)(r).isolatedBacklog}catch(e){}var f={on:p,addEventListener:p,removeEventListener:function(e,t){var r=n[e];if(!r)return;for(var i=0;i<r.length;i++)r[i]===t&&r.splice(i,1)},emit:function(e,r,n,i,o){!1!==o&&(o=!0);if(u.aborted&&!i)return;t&&o&&t.emit(e,r,n);for(var a=g(n),c=m(e),l=c.length,d=0;d<l;d++)c[d].apply(a,r);var p=v()[s[e]];p&&p.push([f,e,r,a]);return a},get:h,listeners:m,context:g,buffer:function(e,t){const r=v();if(t=t||"feature",f.aborted)return;Object.entries(e||{}).forEach((([e,n])=>{s[n]=t,t in r||(r[t]=[])}))},abort:function(){f._aborted=!0,Object.keys(f.backlog).forEach((e=>{delete f.backlog[e]}))},isBuffering:function(e){return!!v()[s[e]]},debugId:r,backlog:d?{}:t&&"object"==typeof t.backlog?t.backlog:{},isolatedBacklog:d};return Object.defineProperty(f,"aborted",{get:()=>{let e=f._aborted||!1;return e||(t&&(e=t.aborted),e)}}),f;function g(e){return e&&e instanceof a.y?e:e?(0,i.I)(e,c,(()=>new a.y(c))):new a.y(c)}function p(e,t){n[e]=m(e).concat(t)}function m(e){return n[e]||[]}function h(t){return l[t]=l[t]||e(f,t)}function v(){return f.backlog}}(void 0,"globalEE"),l=(0,n.Zm)();l.ee||(l.ee=u)},646:(e,t,r)=>{"use strict";r.d(t,{y:()=>n});class n{constructor(e){this.contextId=e}}},908:(e,t,r)=>{"use strict";r.d(t,{d:()=>n,p:()=>i});var n=r(836).ee.get("handle");function i(e,t,r,i,o){o?(o.buffer([e],i),o.emit(e,t,r)):(n.buffer([e],i),n.emit(e,t,r))}},606:(e,t,r)=>{"use strict";r.d(t,{i:()=>o});var n=r(908);o.on=a;var i=o.handlers={};function o(e,t,r,o){a(o||n.d,i,e,t,r)}function a(e,t,r,i,o){o||(o="feature"),e||(e=n.d);var a=t[o]=t[o]||{};(a[r]=a[r]||[]).push([e,i])}},878:(e,t,r)=>{"use strict";function n(e,t){return{capture:e,passive:!1,signal:t}}function i(e,t,r=!1,i){window.addEventListener(e,t,n(r,i))}function o(e,t,r=!1,i){document.addEventListener(e,t,n(r,i))}r.d(t,{DD:()=>o,jT:()=>n,sp:()=>i})},607:(e,t,r)=>{"use strict";r.d(t,{W:()=>n});const n=(0,r(566).bz)()},566:(e,t,r)=>{"use strict";r.d(t,{LA:()=>s,bz:()=>a});var n=r(154);const i="xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx";function o(e,t){return e?15&e[t]:16*Math.random()|0}function a(){const e=n.gm?.crypto||n.gm?.msCrypto;let t,r=0;return e&&e.getRandomValues&&(t=e.getRandomValues(new Uint8Array(30))),i.split("").map((e=>"x"===e?o(t,r++).toString(16):"y"===e?(3&o()|8).toString(16):e)).join("")}function s(e){const t=n.gm?.crypto||n.gm?.msCrypto;let r,i=0;t&&t.getRandomValues&&(r=t.getRandomValues(new Uint8Array(e)));const a=[];for(var s=0;s<e;s++)a.push(o(r,i++).toString(16));return a.join("")}},614:(e,t,r)=>{"use strict";r.d(t,{BB:()=>a,H3:()=>n,g:()=>u,iL:()=>c,tS:()=>s,uh:()=>i,wk:()=>o});const n="NRBA",i="SESSION",o=144e5,a=18e5,s={STARTED:"session-started",PAUSE:"session-pause",RESET:"session-reset",RESUME:"session-resume",UPDATE:"session-update"},c={SAME_TAB:"same-tab",CROSS_TAB:"cross-tab"},u={OFF:0,FULL:1,ERROR:2}},863:(e,t,r)=>{"use strict";function n(){return Math.floor(performance.now())}r.d(t,{t:()=>n})},944:(e,t,r)=>{"use strict";function n(e,t){"function"==typeof console.debug&&console.debug("New Relic Warning: https://github.com/newrelic/newrelic-browser-agent/blob/main/docs/warning-codes.md#".concat(e),t)}r.d(t,{R:()=>n})},284:(e,t,r)=>{"use strict";r.d(t,{t:()=>c,B:()=>s});var n=r(836),i=r(154);const o="newrelic";const a=new Set,s={};function c(e,t){const r=n.ee.get(t);s[t]??={},e&&"object"==typeof e&&(a.has(t)||(r.emit("rumresp",[e]),s[t]=e,a.add(t),function(e={}){try{i.gm.dispatchEvent(new CustomEvent(o,{detail:e}))}catch(e){}}({loaded:!0})))}},990:(e,t,r)=>{"use strict";r.d(t,{I:()=>i});var n=Object.prototype.hasOwnProperty;function i(e,t,r){if(n.call(e,t))return e[t];var i=r();if(Object.defineProperty&&Object.keys)try{return Object.defineProperty(e,t,{value:i,writable:!0,enumerable:!1}),i}catch(e){}return e[t]=i,i}},389:(e,t,r)=>{"use strict";function n(e,t=500,r={}){const n=r?.leading||!1;let i;return(...r)=>{n&&void 0===i&&(e.apply(this,r),i=setTimeout((()=>{i=clearTimeout(i)}),t)),n||(clearTimeout(i),i=setTimeout((()=>{e.apply(this,r)}),t))}}function i(e){let t=!1;return(...r)=>{t||(t=!0,e.apply(this,r))}}r.d(t,{J:()=>i,s:()=>n})},289:(e,t,r)=>{"use strict";r.d(t,{GG:()=>o,sB:()=>a});var n=r(878);function i(){return"undefined"==typeof document||"complete"===document.readyState}function o(e,t){if(i())return e();(0,n.sp)("load",e,t)}function a(e){if(i())return e();(0,n.DD)("DOMContentLoaded",e)}},384:(e,t,r)=>{"use strict";r.d(t,{NT:()=>o,US:()=>l,Zm:()=>a,bQ:()=>c,dV:()=>s,nY:()=>u,pV:()=>d});var n=r(154),i=r(863);const o={beacon:"bam.nr-data.net",errorBeacon:"bam.nr-data.net"};function a(){return n.gm.NREUM||(n.gm.NREUM={}),void 0===n.gm.newrelic&&(n.gm.newrelic=n.gm.NREUM),n.gm.NREUM}function s(){let e=a();return e.o||(e.o={ST:n.gm.setTimeout,SI:n.gm.setImmediate,CT:n.gm.clearTimeout,XHR:n.gm.XMLHttpRequest,REQ:n.gm.Request,EV:n.gm.Event,PR:n.gm.Promise,MO:n.gm.MutationObserver,FETCH:n.gm.fetch,WS:n.gm.WebSocket}),e}function c(e,t){let r=a();r.initializedAgents??={},t.initializedAt={ms:(0,i.t)(),date:new Date},r.initializedAgents[e]=t}function u(e){let t=a();return t.initializedAgents?.[e]}function l(e,t){a()[e]=t}function d(){return function(){let e=a();const t=e.info||{};e.info={beacon:o.beacon,errorBeacon:o.errorBeacon,...t}}(),function(){let e=a();const t=e.init||{};e.init={...t}}(),s(),function(){let e=a();const t=e.loader_config||{};e.loader_config={...t}}(),a()}},843:(e,t,r)=>{"use strict";r.d(t,{u:()=>i});var n=r(878);function i(e,t=!1,r,i){(0,n.DD)("visibilitychange",(function(){if(t)return void("hidden"===document.visibilityState&&e());e(document.visibilityState)}),r,i)}},434:(e,t,r)=>{"use strict";r.d(t,{Jt:()=>o,YM:()=>c});var n=r(836),i=r(607);const o="nr@original:".concat(i.W);var a=Object.prototype.hasOwnProperty,s=!1;function c(e,t){return e||(e=n.ee),r.inPlace=function(e,t,n,i,o){n||(n="");const a="-"===n.charAt(0);for(let s=0;s<t.length;s++){const c=t[s],u=e[c];l(u)||(e[c]=r(u,a?c+n:n,i,c,o))}},r.flag=o,r;function r(t,r,n,s,c){return l(t)?t:(r||(r=""),nrWrapper[o]=t,function(e,t,r){if(Object.defineProperty&&Object.keys)try{return Object.keys(e).forEach((function(r){Object.defineProperty(t,r,{get:function(){return e[r]},set:function(t){return e[r]=t,t}})})),t}catch(e){u([e],r)}for(var n in e)a.call(e,n)&&(t[n]=e[n])}(t,nrWrapper,e),nrWrapper);function nrWrapper(){var o,a,l,d;try{a=this,o=[...arguments],l="function"==typeof n?n(o,a):n||{}}catch(t){u([t,"",[o,a,s],l],e)}i(r+"start",[o,a,s],l,c);try{return d=t.apply(a,o)}catch(e){throw i(r+"err",[o,a,e],l,c),e}finally{i(r+"end",[o,a,d],l,c)}}}function i(r,n,i,o){if(!s||t){var a=s;s=!0;try{e.emit(r,n,i,t,o)}catch(t){u([t,r,n,i],e)}s=a}}}function u(e,t){t||(t=n.ee);try{t.emit("internal-error",e)}catch(e){}}function l(e){return!(e&&"function"==typeof e&&e.apply&&!e[o])}},993:(e,t,r)=>{"use strict";r.d(t,{A$:()=>o,ET:()=>a,p_:()=>i});var n=r(860);const i={ERROR:"ERROR",WARN:"WARN",INFO:"INFO",DEBUG:"DEBUG",TRACE:"TRACE"},o={OFF:0,ERROR:1,WARN:2,INFO:3,DEBUG:4,TRACE:5},a="log";n.K7.logging},773:(e,t,r)=>{"use strict";r.d(t,{z_:()=>o,XG:()=>s,TZ:()=>n,rs:()=>i,xV:()=>a});r(154),r(566),r(384);const n=r(860).K7.metrics,i="sm",o="cm",a="storeSupportabilityMetrics",s="storeEventMetrics"},630:(e,t,r)=>{"use strict";r.d(t,{T:()=>n});const n=r(860).K7.pageViewEvent},782:(e,t,r)=>{"use strict";r.d(t,{T:()=>n});const n=r(860).K7.pageViewTiming},344:(e,t,r)=>{"use strict";r.d(t,{G4:()=>i});var n=r(614);r(860).K7.sessionReplay;const i={RECORD:"recordReplay",PAUSE:"pauseReplay",REPLAY_RUNNING:"replayRunning",ERROR_DURING_REPLAY:"errorDuringReplay"};n.g.ERROR,n.g.FULL,n.g.OFF},234:(e,t,r)=>{"use strict";r.d(t,{W:()=>o});var n=r(836),i=r(687);class o{constructor(e,t){this.agentIdentifier=e,this.ee=n.ee.get(e),this.featureName=t,this.blocked=!1}deregisterDrain(){(0,i.x3)(this.agentIdentifier,this.featureName)}}},603:(e,t,r)=>{"use strict";r.d(t,{j:()=>K});var n=r(860),i=r(555),o=r(371),a=r(908),s=r(836),c=r(687),u=r(289),l=r(154),d=r(944),f=r(773),g=r(384),p=r(344);const m=["setErrorHandler","finished","addToTrace","addRelease","recordCustomEvent","addPageAction","setCurrentRouteName","setPageViewName","setCustomAttribute","interaction","noticeError","setUserId","setApplicationVersion","start",p.G4.RECORD,p.G4.PAUSE,"log","wrapLogger"],h=["setErrorHandler","finished","addToTrace","addRelease"];var v=r(863),b=r(614),y=r(993);var w=r(646),R=r(434);const A=new Map;function E(e,t,r,n){if("object"!=typeof t||!t||"string"!=typeof r||!r||"function"!=typeof t[r])return(0,d.R)(29);const i=function(e){return(e||s.ee).get("logger")}(e),o=(0,R.YM)(i),a=new w.y(s.P);a.level=n.level,a.customAttributes=n.customAttributes;const c=t[r]?.[R.Jt]||t[r];return A.set(c,a),o.inPlace(t,[r],"wrap-logger-",(()=>A.get(c))),i}function _(){const e=(0,g.pV)();m.forEach((t=>{e[t]=(...r)=>function(t,...r){let n=[];return Object.values(e.initializedAgents).forEach((e=>{e&&e.api?e.exposed&&e.api[t]&&n.push(e.api[t](...r)):(0,d.R)(38,t)})),n.length>1?n:n[0]}(t,...r)}))}const x={};function N(e,t,g=!1){t||(0,c.Ak)(e,"api");const m={};var w=s.ee.get(e),R=w.get("tracer");x[e]=b.g.OFF,w.on(p.G4.REPLAY_RUNNING,(t=>{x[e]=t}));var A="api-",_=A+"ixn-";function N(t,r,n,o){const a=(0,i.Vp)(e);return null===r?delete a.jsAttributes[t]:(0,i.x1)(e,{...a,jsAttributes:{...a.jsAttributes,[t]:r}}),j(A,n,!0,o||null===r?"session":void 0)(t,r)}function k(){}m.log=function(e,{customAttributes:t={},level:r=y.p_.INFO}={}){(0,a.p)(f.xV,["API/log/called"],void 0,n.K7.metrics,w),function(e,t,r={},i=y.p_.INFO){(0,a.p)(f.xV,["API/logging/".concat(i.toLowerCase(),"/called")],void 0,n.K7.metrics,e),(0,a.p)(y.ET,[(0,v.t)(),t,r,i],void 0,n.K7.logging,e)}(w,e,t,r)},m.wrapLogger=(e,t,{customAttributes:r={},level:i=y.p_.INFO}={})=>{(0,a.p)(f.xV,["API/wrapLogger/called"],void 0,n.K7.metrics,w),E(w,e,t,{customAttributes:r,level:i})},h.forEach((e=>{m[e]=j(A,e,!0,"api")})),m.addPageAction=j(A,"addPageAction",!0,n.K7.genericEvents),m.recordCustomEvent=j(A,"recordCustomEvent",!0,n.K7.genericEvents),m.setPageViewName=function(t,r){if("string"==typeof t)return"/"!==t.charAt(0)&&(t="/"+t),(0,o.f)(e).customTransaction=(r||"http://custom.transaction")+t,j(A,"setPageViewName",!0)()},m.setCustomAttribute=function(e,t,r=!1){if("string"==typeof e){if(["string","number","boolean"].includes(typeof t)||null===t)return N(e,t,"setCustomAttribute",r);(0,d.R)(40,typeof t)}else(0,d.R)(39,typeof e)},m.setUserId=function(e){if("string"==typeof e||null===e)return N("enduser.id",e,"setUserId",!0);(0,d.R)(41,typeof e)},m.setApplicationVersion=function(e){if("string"==typeof e||null===e)return N("application.version",e,"setApplicationVersion",!1);(0,d.R)(42,typeof e)},m.start=()=>{try{(0,a.p)(f.xV,["API/start/called"],void 0,n.K7.metrics,w),w.emit("manual-start-all")}catch(e){(0,d.R)(23,e)}},m[p.G4.RECORD]=function(){(0,a.p)(f.xV,["API/recordReplay/called"],void 0,n.K7.metrics,w),(0,a.p)(p.G4.RECORD,[],void 0,n.K7.sessionReplay,w)},m[p.G4.PAUSE]=function(){(0,a.p)(f.xV,["API/pauseReplay/called"],void 0,n.K7.metrics,w),(0,a.p)(p.G4.PAUSE,[],void 0,n.K7.sessionReplay,w)},m.interaction=function(e){return(new k).get("object"==typeof e?e:{})};const T=k.prototype={createTracer:function(e,t){var r={},i=this,o="function"==typeof t;return(0,a.p)(f.xV,["API/createTracer/called"],void 0,n.K7.metrics,w),g||(0,a.p)(_+"tracer",[(0,v.t)(),e,r],i,n.K7.spa,w),function(){if(R.emit((o?"":"no-")+"fn-start",[(0,v.t)(),i,o],r),o)try{return t.apply(this,arguments)}catch(e){const t="string"==typeof e?new Error(e):e;throw R.emit("fn-err",[arguments,this,t],r),t}finally{R.emit("fn-end",[(0,v.t)()],r)}}}};function j(e,t,r,i){return function(){return(0,a.p)(f.xV,["API/"+t+"/called"],void 0,n.K7.metrics,w),i&&(0,a.p)(e+t,[r?(0,v.t)():performance.now(),...arguments],r?null:this,i,w),r?void 0:this}}function I(){r.e(296).then(r.bind(r,778)).then((({setAPI:t})=>{t(e),(0,c.Ze)(e,"api")})).catch((e=>{(0,d.R)(27,e),w.abort()}))}return["actionText","setName","setAttribute","save","ignore","onEnd","getContext","end","get"].forEach((e=>{T[e]=j(_,e,void 0,g?n.K7.softNav:n.K7.spa)})),m.setCurrentRouteName=g?j(_,"routeName",void 0,n.K7.softNav):j(A,"routeName",!0,n.K7.spa),m.noticeError=function(t,r){"string"==typeof t&&(t=new Error(t)),(0,a.p)(f.xV,["API/noticeError/called"],void 0,n.K7.metrics,w),(0,a.p)("err",[t,(0,v.t)(),!1,r,!!x[e]],void 0,n.K7.jserrors,w)},l.RI?(0,u.GG)((()=>I()),!0):I(),m}var k=r(217),T=r(122);const j={accountID:void 0,trustKey:void 0,agentID:void 0,licenseKey:void 0,applicationID:void 0,xpid:void 0},I={};var O=r(284);const S=e=>{const t=e.startsWith("http");e+="/",r.p=t?e:"https://"+e};let P=!1;function K(e,t={},r,n){let{init:a,info:c,loader_config:u,runtime:d={},exposed:f=!0}=t;d.loaderType=r;const p=(0,g.pV)();c||(a=p.init,c=p.info,u=p.loader_config),(0,k.xN)(e.agentIdentifier,a||{}),function(e,t){if(!e)throw new Error("All loader-config objects require an agent identifier!");I[e]=(0,T.a)(t,j);const r=(0,g.nY)(e);r&&(r.loader_config=I[e])}(e.agentIdentifier,u||{}),c.jsAttributes??={},l.bv&&(c.jsAttributes.isWorker=!0),(0,i.x1)(e.agentIdentifier,c);const m=(0,k.D0)(e.agentIdentifier),h=[c.beacon,c.errorBeacon];P||(m.proxy.assets&&(S(m.proxy.assets),h.push(m.proxy.assets)),m.proxy.beacon&&h.push(m.proxy.beacon),_(),(0,g.US)("activatedFeatures",O.B),e.runSoftNavOverSpa&&=!0===m.soft_navigations.enabled&&m.feature_flags.includes("soft_nav")),d.denyList=[...m.ajax.deny_list||[],...m.ajax.block_internal?h:[]],d.ptid=e.agentIdentifier,(0,o.V)(e.agentIdentifier,d),e.ee=s.ee.get(e.agentIdentifier),void 0===e.api&&(e.api=N(e.agentIdentifier,n,e.runSoftNavOverSpa)),void 0===e.exposed&&(e.exposed=f),P=!0}},374:(e,t,r)=>{r.nc=(()=>{try{return document?.currentScript?.nonce}catch(e){}return""})()},860:(e,t,r)=>{"use strict";r.d(t,{$J:()=>u,K7:()=>s,P3:()=>c,XX:()=>i,qY:()=>n,v4:()=>a});const n="events",i="jserrors",o="browser/blobs",a="rum",s={ajax:"ajax",genericEvents:"generic_events",jserrors:i,logging:"logging",metrics:"metrics",pageAction:"page_action",pageViewEvent:"page_view_event",pageViewTiming:"page_view_timing",sessionReplay:"session_replay",sessionTrace:"session_trace",softNav:"soft_navigations",spa:"spa"},c={[s.pageViewEvent]:1,[s.pageViewTiming]:2,[s.metrics]:3,[s.jserrors]:4,[s.spa]:5,[s.ajax]:6,[s.sessionTrace]:7,[s.softNav]:8,[s.sessionReplay]:9,[s.logging]:10,[s.genericEvents]:11},u={[s.pageViewEvent]:a,[s.pageViewTiming]:n,[s.ajax]:n,[s.spa]:n,[s.softNav]:n,[s.metrics]:i,[s.jserrors]:i,[s.sessionTrace]:o,[s.sessionReplay]:o,[s.logging]:"browser/logs",[s.genericEvents]:"ins"}}},n={};function i(e){var t=n[e];if(void 0!==t)return t.exports;var o=n[e]={exports:{}};return r[e](o,o.exports,i),o.exports}i.m=r,i.d=(e,t)=>{for(var r in t)i.o(t,r)&&!i.o(e,r)&&Object.defineProperty(e,r,{enumerable:!0,get:t[r]})},i.f={},i.e=e=>Promise.all(Object.keys(i.f).reduce(((t,r)=>(i.f[r](e,t),t)),[])),i.u=e=>"nr-rum-1.283.2.min.js",i.o=(e,t)=>Object.prototype.hasOwnProperty.call(e,t),e={},t="NRBA-1.283.2.PROD:",i.l=(r,n,o,a)=>{if(e[r])e[r].push(n);else{var s,c;if(void 0!==o)for(var u=document.getElementsByTagName("script"),l=0;l<u.length;l++){var d=u[l];if(d.getAttribute("src")==r||d.getAttribute("data-webpack")==t+o){s=d;break}}if(!s){c=!0;var f={296:"sha512-2Y8GMAOGF658KnXzOZ/v+DlLch8TBFvV0tTNnOy9wrpvtDa1t5CdZMyX+LubTymBlzPp6NUjllBghMCZqXBPmg=="};(s=document.createElement("script")).charset="utf-8",s.timeout=120,i.nc&&s.setAttribute("nonce",i.nc),s.setAttribute("data-webpack",t+o),s.src=r,0!==s.src.indexOf(window.location.origin+"/")&&(s.crossOrigin="anonymous"),f[a]&&(s.integrity=f[a])}e[r]=[n];var g=(t,n)=>{s.onerror=s.onload=null,clearTimeout(p);var i=e[r];if(delete e[r],s.parentNode&&s.parentNode.removeChild(s),i&&i.forEach((e=>e(n))),t)return t(n)},p=setTimeout(g.bind(null,void 0,{type:"timeout",target:s}),12e4);s.onerror=g.bind(null,s.onerror),s.onload=g.bind(null,s.onload),c&&document.head.appendChild(s)}},i.r=e=>{"undefined"!=typeof Symbol&&Symbol.toStringTag&&Object.defineProperty(e,Symbol.toStringTag,{value:"Module"}),Object.defineProperty(e,"__esModule",{value:!0})},i.p="https://js-agent.newrelic.com/",(()=>{var e={374:0,840:0};i.f.j=(t,r)=>{var n=i.o(e,t)?e[t]:void 0;if(0!==n)if(n)r.push(n[2]);else{var o=new Promise(((r,i)=>n=e[t]=[r,i]));r.push(n[2]=o);var a=i.p+i.u(t),s=new Error;i.l(a,(r=>{if(i.o(e,t)&&(0!==(n=e[t])&&(e[t]=void 0),n)){var o=r&&("load"===r.type?"missing":r.type),a=r&&r.target&&r.target.src;s.message="Loading chunk "+t+" failed.\n("+o+": "+a+")",s.name="ChunkLoadError",s.type=o,s.request=a,n[1](s)}}),"chunk-"+t,t)}};var t=(t,r)=>{var n,o,[a,s,c]=r,u=0;if(a.some((t=>0!==e[t]))){for(n in s)i.o(s,n)&&(i.m[n]=s[n]);if(c)c(i)}for(t&&t(r);u<a.length;u++)o=a[u],i.o(e,o)&&e[o]&&e[o][0](),e[o]=0},r=self["webpackChunk:NRBA-1.283.2.PROD"]=self["webpackChunk:NRBA-1.283.2.PROD"]||[];r.forEach(t.bind(null,0)),r.push=t.bind(null,r.push.bind(r))})(),(()=>{"use strict";i(374);var e=i(944),t=i(344),r=i(566);class n{agentIdentifier;constructor(){this.agentIdentifier=(0,r.LA)(16)}#e(t,...r){if("function"==typeof this.api?.[t])return this.api[t](...r);(0,e.R)(35,t)}addPageAction(e,t){return this.#e("addPageAction",e,t)}recordCustomEvent(e,t){return this.#e("recordCustomEvent",e,t)}setPageViewName(e,t){return this.#e("setPageViewName",e,t)}setCustomAttribute(e,t,r){return this.#e("setCustomAttribute",e,t,r)}noticeError(e,t){return this.#e("noticeError",e,t)}setUserId(e){return this.#e("setUserId",e)}setApplicationVersion(e){return this.#e("setApplicationVersion",e)}setErrorHandler(e){return this.#e("setErrorHandler",e)}addRelease(e,t){return this.#e("addRelease",e,t)}log(e,t){return this.#e("log",e,t)}}class o extends n{#e(t,...r){if("function"==typeof this.api?.[t])return this.api[t](...r);(0,e.R)(35,t)}start(){return this.#e("start")}finished(e){return this.#e("finished",e)}recordReplay(){return this.#e(t.G4.RECORD)}pauseReplay(){return this.#e(t.G4.PAUSE)}addToTrace(e){return this.#e("addToTrace",e)}setCurrentRouteName(e){return this.#e("setCurrentRouteName",e)}interaction(){return this.#e("interaction")}wrapLogger(e,t,r){return this.#e("wrapLogger",e,t,r)}}var a=i(860),s=i(217);const c=Object.values(a.K7);function u(e){const t={};return c.forEach((r=>{t[r]=function(e,t){return!0===(0,s.gD)(t,"".concat(e,".enabled"))}(r,e)})),t}var l=i(603);var d=i(687),f=i(234),g=i(289),p=i(154),m=i(384);const h=e=>p.RI&&!0===(0,s.gD)(e,"privacy.cookies_enabled");function v(e){return!!(0,m.dV)().o.MO&&h(e)&&!0===(0,s.gD)(e,"session_trace.enabled")}var b=i(389);class y extends f.W{constructor(e,t,r=!0){super(e.agentIdentifier,t),this.auto=r,this.abortHandler=void 0,this.featAggregate=void 0,this.onAggregateImported=void 0,!1===e.init[this.featureName].autoStart&&(this.auto=!1),this.auto?(0,d.Ak)(e.agentIdentifier,t):this.ee.on("manual-start-all",(0,b.J)((()=>{(0,d.Ak)(e.agentIdentifier,this.featureName),this.auto=!0,this.importAggregator(e)})))}importAggregator(t,r={}){if(this.featAggregate||!this.auto)return;let n;this.onAggregateImported=new Promise((e=>{n=e}));const o=async()=>{let o;try{if(h(this.agentIdentifier)){const{setupAgentSession:e}=await i.e(296).then(i.bind(i,861));o=e(t)}}catch(t){(0,e.R)(20,t),this.ee.emit("internal-error",[t]),this.featureName===a.K7.sessionReplay&&this.abortHandler?.()}try{if(!this.#t(this.featureName,o))return(0,d.Ze)(this.agentIdentifier,this.featureName),void n(!1);const{lazyFeatureLoader:e}=await i.e(296).then(i.bind(i,103)),{Aggregate:a}=await e(this.featureName,"aggregate");this.featAggregate=new a(t,r),t.runtime.harvester.initializedAggregates.push(this.featAggregate),n(!0)}catch(t){(0,e.R)(34,t),this.abortHandler?.(),(0,d.Ze)(this.agentIdentifier,this.featureName,!0),n(!1),this.ee&&this.ee.abort()}};p.RI?(0,g.GG)((()=>o()),!0):o()}#t(e,t){switch(e){case a.K7.sessionReplay:return v(this.agentIdentifier)&&!!t;case a.K7.sessionTrace:return!!t;default:return!0}}}var w=i(630);class R extends y{static featureName=w.T;constructor(e,t=!0){super(e,w.T,t),this.importAggregator(e)}}var A=i(908),E=i(843),_=i(878),x=i(782),N=i(863);class k extends y{static featureName=x.T;constructor(e,t=!0){super(e,x.T,t),p.RI&&((0,E.u)((()=>(0,A.p)("docHidden",[(0,N.t)()],void 0,x.T,this.ee)),!0),(0,_.sp)("pagehide",(()=>(0,A.p)("winPagehide",[(0,N.t)()],void 0,x.T,this.ee))),this.importAggregator(e))}}var T=i(773);class j extends y{static featureName=T.TZ;constructor(e,t=!0){super(e,T.TZ,t),this.importAggregator(e)}}new class extends o{constructor(t){super(),p.gm?(this.features={},(0,m.bQ)(this.agentIdentifier,this),this.desiredFeatures=new Set(t.features||[]),this.desiredFeatures.add(R),this.runSoftNavOverSpa=[...this.desiredFeatures].some((e=>e.featureName===a.K7.softNav)),(0,l.j)(this,t,t.loaderType||"agent"),this.run()):(0,e.R)(21)}get config(){return{info:this.info,init:this.init,loader_config:this.loader_config,runtime:this.runtime}}run(){try{const t=u(this.agentIdentifier),r=[...this.desiredFeatures];r.sort(((e,t)=>a.P3[e.featureName]-a.P3[t.featureName])),r.forEach((r=>{if(!t[r.featureName]&&r.featureName!==a.K7.pageViewEvent)return;if(this.runSoftNavOverSpa&&r.featureName===a.K7.spa)return;if(!this.runSoftNavOverSpa&&r.featureName===a.K7.softNav)return;const n=function(e){switch(e){case a.K7.ajax:return[a.K7.jserrors];case a.K7.sessionTrace:return[a.K7.ajax,a.K7.pageViewEvent];case a.K7.sessionReplay:return[a.K7.sessionTrace];case a.K7.pageViewTiming:return[a.K7.pageViewEvent];default:return[]}}(r.featureName).filter((e=>!(e in this.features)));n.length>0&&(0,e.R)(36,{targetFeature:r.featureName,missingDependencies:n}),this.features[r.featureName]=new r(this)}))}catch(t){(0,e.R)(22,t);for(const e in this.features)this.features[e].abortHandler?.();const r=(0,m.Zm)();delete r.initializedAgents[this.agentIdentifier]?.api,delete r.initializedAgents[this.agentIdentifier]?.features,delete this.sharedAggregator;return r.ee.get(this.agentIdentifier).abort(),!1}}}({features:[R,k,j],loaderType:"lite"})})()})();</script> <meta name="viewport" content="width=device-width, initial-scale=1, shrink-to-fit=no"> <link rel="shortcut icon" href="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/dist/images/favicon_300a1064.ico" type="image/vnd.microsoft.icon"> <link media="all" href="https://developer-blogs.nvidia.com/wp-content/uploads/autoptimize/1/css/aggregated_1cf4c58a298ccc92d72f18db44a919e3.css" rel="stylesheet"><title>Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO | NVIDIA Technical Blog</title>  <script src="https://cdn.cookielaw.org/scripttemplates/otSDKStub.js" data-document-language="true" type="text/javascript" charset="UTF-8" data-domain-script="3e2b62ff-7ae7-4ac5-87c8-d5949ecafff5" ></script> <script type="text/javascript"> function OptanonWrapper() { var event = new Event('bannerLoaded'); window.dispatchEvent(event); } </script> <link rel="preload" href="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/dist/scripts/main_61b387dd.js" as="script"><script src="//app-sj14.marketo.com/js/forms2/js/forms2.min.js?ver=5.6.2" defer></script><script src="https://unpkg.com/infinite-scroll@4/dist/infinite-scroll.pkgd.min.js" defer></script><meta property="interest" content="Computer Vision / Video Analytics"><meta property="estimated-read-time" content="14 MIN">  <meta name="robots" content="max-snippet:-1,max-image-preview:standard,max-video-preview:-1" /> <link rel="canonical" href="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/" /> <link rel="shortlink" href="https://developer.nvidia.com/blog/?p=68713" /> <meta name="description" content="Learn how to use NVIDIA TAO and pretrained models to create and deploy custom AI models to accurately detect and recognize handwritten texts." /> <meta property="og:type" content="article" /> <meta property="og:locale" content="en_US" /> <meta property="og:site_name" content="NVIDIA Technical Blog" /> <meta property="og:title" content="Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO | NVIDIA Technical Blog" /> <meta property="og:description" content="Learn how to use NVIDIA TAO and pretrained models to create and deploy custom AI models to accurately detect and recognize handwritten texts." /> <meta property="og:url" content="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/" /> <meta property="og:image" content="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail.png" /> <meta property="og:image:width" content="1920" /> <meta property="og:image:height" content="1080" /> <meta property="article:published_time" content="2023-08-15T15:00+00:00" /> <meta property="article:modified_time" content="2023-08-24T18:03+00:00" /> <meta name="twitter:card" content="summary_large_image" /> <meta name="twitter:title" content="Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO | NVIDIA Technical Blog" /> <meta name="twitter:description" content="Learn how to use NVIDIA TAO and pretrained models to create and deploy custom AI models to accurately detect and recognize handwritten texts." /> <meta name="twitter:image" content="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail.png" /> <script type="application/ld+json">{"@context":"https://schema.org","@graph":[{"@type":"WebSite","@id":"https://developer.nvidia.com/blog/#/schema/WebSite","url":"https://developer.nvidia.com/blog/","name":"NVIDIA Technical Blog","description":"News and tutorials for developers, data scientists, and IT admins","inLanguage":"en-US","potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://developer.nvidia.com/blog/search/{search_term_string}/"},"query-input":"required name=search_term_string"},"publisher":{"@type":"Organization","@id":"https://developer.nvidia.com/blog/#/schema/Organization","name":"NVIDIA Technical Blog","url":"https://developer.nvidia.com/blog/"}},{"@type":"WebPage","@id":"https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/","url":"https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/","name":"Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO | NVIDIA Technical Blog","description":"Learn how to use NVIDIA TAO and pretrained models to create and deploy custom AI models to accurately detect and recognize handwritten texts.","inLanguage":"en-US","isPartOf":{"@id":"https://developer.nvidia.com/blog/#/schema/WebSite"},"breadcrumb":{"@type":"BreadcrumbList","@id":"https://developer.nvidia.com/blog/#/schema/BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"item":"https://developer.nvidia.com/blog/","name":"NVIDIA Technical Blog"},{"@type":"ListItem","position":2,"item":"https://developer.nvidia.com/blog/category/computer-vision/","name":"Category: Computer Vision / Video Analytics"},{"@type":"ListItem","position":3,"name":"Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO"}]},"potentialAction":{"@type":"ReadAction","target":"https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/"},"datePublished":"2023-08-15T15:00+00:00","dateModified":"2023-08-24T18:03+00:00","author":{"@type":"Person","@id":"https://developer.nvidia.com/blog/#/schema/Person/89d3a6133aa6b238cf7b6b0b16462e33","name":"Chintan Shah","description":"Chintan Shah is a senior product manager at NVIDIA, focusing on AI products. He manages NVIDIA TAO and vision AI solutions for smart spaces, retail, industrial, and other industry verticals. His focus is on simplifying and democratizing AI for all..."}}]}</script>  <link rel='dns-prefetch' href='//secure.gravatar.com' /> <link rel='dns-prefetch' href='//widgets.wp.com' /> <link rel='dns-prefetch' href='//s0.wp.com' /> <link rel='dns-prefetch' href='//0.gravatar.com' /> <link rel='dns-prefetch' href='//1.gravatar.com' /> <link rel='dns-prefetch' href='//2.gravatar.com' /> <link rel='dns-prefetch' href='//v0.wordpress.com' /> <link href='https://assets.adobedtm.com' rel='preconnect' /> <link rel="alternate" type="application/atom+xml" title="NVIDIA Technical Blog » Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO Comments Feed" href="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/feed/" /> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/resources/assets/scripts/munchkin.js?ver=1740539867" id="munchkin-js-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/resources/assets/scripts/user-dropdown.js?ver=1740539867" id="user_dropdown.js-js"></script> <link rel="https://api.w.org/" href="https://developer-blogs.nvidia.com/wp-json/" /><link rel="alternate" title="JSON" type="application/json" href="https://developer-blogs.nvidia.com/wp-json/wp/v2/posts/68713" /><link rel="EditURI" type="application/rsd+xml" title="RSD" href="https://developer-blogs.nvidia.com/xmlrpc.php?rsd" /> <link rel="alternate" title="oEmbed (JSON)" type="application/json+oembed" href="https://developer-blogs.nvidia.com/wp-json/oembed/1.0/embed?url=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" /> <link rel="alternate" title="oEmbed (XML)" type="text/xml+oembed" href="https://developer-blogs.nvidia.com/wp-json/oembed/1.0/embed?url=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F&format=xml" /> <link rel="alternate" hreflang="en" href="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/" title="en_US" /> <link rel="alternate" hreflang="zh" href="https://developer.nvidia.com/zh-cn/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/" title="zh_CN" />   <script src="//assets.adobedtm.com/5d4962a43b79/c1061d2c5e7b/launch-191c2462b890.min.js" data-ot-ignore="true"></script> <link rel="stylesheet" href="https://pro.fontawesome.com/releases/v5.12.1/css/all.css" integrity="sha384-TxKWSXbsweFt0o2WqfkfJRRNVaPdzXJ/YLqgStggBVRREXkwU7OKz+xXtqOU4u8+" crossorigin="anonymous" media="none" onload="if(media!='all')media='all'"> <script src='https://developer-blogs.nvidia.com/wp-content/plugins/nv-faceted-search/js/faceted-search/build/direflowBundle.js?ver=6daec84'></script> </head> <body class="post-template-default single single-post postid-68713 single-format-standard create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1 app-data index-data singular-data single-data single-post-data single-post-create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1-data"> <svg xmlns="http://www.w3.org/2000/svg" class="hide" style="display: none;"> <symbol id="n24-user-circle" viewBox="0 0 24 24"> <defs> </defs> <rect id="n24-icon-user-circle-bounds" class="n24-bounds" width="24" height="24"></rect> <g id="n24-icon-user-circle"> <circle id="user-circle" class="n24-user-circle-cls-1" cx="12" cy="12" r="9.75"></circle> <polyline id="user-body" class="n24-user-circle-cls-1" points="17.55 20.01 16.34 15.48 7.66 15.48 6.45 20.01"></polyline> <circle id="user-head" class="n24-user-circle-cls-1" cx="12" cy="9.91" r="3.48"></circle> </g> </symbol> </svg> <header class="page-header"> <div class="container"> <button class="navbar-toggle" type="button"> <span class="icon-bar"></span> <span class="icon-bar"></span> <span class="icon-bar"></span> </button> <div class="logo"> <a class="primary-logo" href="/" title="Home" data-wpel-link="internal" target="_blank" rel="follow noopener"> <img alt="Home" width="142" height="32" src="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/dist/images/nvidia-logo_28b633c7.svg"> </a> <a class="secondary-logo" href="/" title="Home" data-wpel-link="internal" target="_blank" rel="follow noopener">DEVELOPER</a> </div> <nav class="navbar primary navbar-expand-md navbar-inverse"> <div class="collapse navbar-collapse" id="navbarSupportedContent"> <div class="menu-top-menu-container"><ul id="menu-top-menu" class="menu nav navbar-nav primary mr-auto"><li id="menu-item-25291" class="menu-item menu-item-type-custom menu-item-object-custom nav-item nav-item-25291"><a title="Home" href="/" class="nav-link" data-wpel-link="internal" target="_blank" rel="follow noopener">Home</a></li><li id="menu-item-25292" class="menu-item menu-item-type-custom menu-item-object-custom nav-item nav-item-25292"><a title="Blog" href="/blog" class="nav-link active" data-wpel-link="internal" target="_blank" rel="follow noopener">Blog</a></li><li id="menu-item-25293" class="menu-item menu-item-type-custom menu-item-object-custom nav-item nav-item-25293"><a title="Forums" href="https://forums.developer.nvidia.com/" class="nav-link" data-wpel-link="internal" target="_blank" rel="follow noopener">Forums</a></li><li id="menu-item-25294" class="menu-item menu-item-type-custom menu-item-object-custom nav-item nav-item-25294"><a title="Docs" href="https://docs.nvidia.com/" class="nav-link" data-wpel-link="internal" target="_blank" rel="follow noopener">Docs</a></li><li id="menu-item-25295" class="menu-item menu-item-type-custom menu-item-object-custom nav-item nav-item-25295"><a title="Downloads" href="https://developer.nvidia.com/downloads" class="nav-link" data-wpel-link="internal" target="_blank" rel="follow noopener">Downloads</a></li><li id="menu-item-25296" class="menu-item menu-item-type-custom menu-item-object-custom nav-item nav-item-25296"><a title="Training" href="https://www.nvidia.com/en-us/training/" class="nav-link" data-wpel-link="internal" target="_blank" rel="follow noopener">Training</a></li></ul></div> </div> </nav> <nav class="navbar navbar-right" id="topRightNavbar"> <ul class="nav navbar-nav navbar-margin navbar-right navbar-margin-media login-nav"> <li class="search" id="search-top"> <div class="search-form" id="search-top-form"> <form class="gss form-search content-search" data-base-url="https://developer.nvidia.com/blog/" action="https://developer.nvidia.com/blog/search-posts" method="GET" id="nvidia-site-search-form" accept-charset="UTF-8"> <div> <div class="input-group"> <div id="top-site-search-form"> <input placeholder="Search" class="form-control form-text" type="text" id="edit-term" name="q" value="" size="15" maxlength="128"> </div> <span class="input-group-btn"> <button type="submit" class="btn btn-default"> <i class="fa fa-search" aria-hidden="true"></i> </button> </span> </div> <button class="btn element-invisible btn-primary form-submit" type="submit" id="edit-submit" name="op" value="Search">Search</button> </div> </form> </div> </li> <li class="leaf" id="dzauth_join_link"> <a href="https://developer.nvidia.com/login" class="cta--prim nv-join" data-wpel-link="internal" target="_blank" rel="follow noopener">Join</a> </li> <li class="leaf last" id="dzauth_login_link"> <a href="https://developer.nvidia.com/login" class="nv-login" data-wpel-link="internal" target="_blank" rel="follow noopener"> <svg aria-label="User Account" class="n24-icon nvprofileicon n24-account" style="width: 24px; height: 24px;"> <use xlink:href="#n24-user-circle"></use> </svg> </a> </li> </ul> </nav> </div> </header> <section class="main"> <div class="faceted-search-top-wrapper"> <div class="container"> <div class="row"> <div class="col col-lg-12 faceted-search-component-wrapper"> <faceted-search-bar appUrl="https://developer.nvidia.com/blog" componentTitle="Technical Blog" text="" subscribeUrl="https://developer.nvidia.com/email-signup" subscribeLabel="Subscribe" locale="en-US" redirectUrl="https://developer.nvidia.com/blog/search-posts/" autocompleteEndpointUrl="https://developer.nvidia.com/blog/wp-json/faceted-search/v1/autocomplete/" data-search-page="false"></faceted-search-bar> </div> </div> </div> </div> <div class="search-filters-component-wrapper--inner-page position-fixed w-100"> <div class="container"> <div class="row"> <faceted-search-filters endpoint="https://developer.nvidia.com/blog/wp-json/faceted-search/v1/select/" redirectUrl="https://developer.nvidia.com/blog/search-posts/" data-expanded="false" data-with-redirect="true" data-with-overlay="false" sortBy="post_date" locale="en-US" handleURls=""></faceted-search-filters> </div> </div> </div> <div class="container single-page"> <div id="main-content" class="row"> <script> sessionStorage.removeItem('facets.search.text'); sessionStorage.removeItem('facets.search.text-copy'); sessionStorage.removeItem('facets.search.filters'); </script> <div class="col-sm-12 d-none text-right sidebar-toggle-link-wrapper"> <a class="sidebar-toggle sidebar-toggle-link" href="#main-content-end"> Related Resources <i class="fa fa-angle-right" aria-hidden="true"></i> </a> </div> <div class="main-content col-lg-9 col-md-9 mt-0 mb-0"> <div class="post-card--single"> <div class="row"> <div class="col-lg-12 mb-0"> <div class="card--post-attributes"> <span class="category-name content-s"> <a href="https://developer.nvidia.com/blog/category/computer-vision/" data-wpel-link="internal" target="_blank" rel="follow noopener">Computer Vision / Video Analytics</a> </span> <span class="post-rate-widget content-s"></span> <span class="post-lang-switcher"> <div class="posts-filter-form"><div class="filter-item form-type-select"><select class="msls_languages"><option value="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/" selected>English</option><option value="https://developer.nvidia.com/zh-cn/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1/">中文</option></select></div></div> </span> </div> <h1 class="h--large txt-clr--blck mt-2 mb-0">Customizing AI Models: Train Character Detection and Recognition Models with NVIDIA TAO</h1> <div class="post-info"> <div class="post-published-date"> Aug 15, 2023 </div> <div class="post-authors"> By <a href="https://developer.nvidia.com/blog/author/chintan-shah/" title="Posts by Chintan Shah" class="author url fn" rel="author follow noopener" data-wpel-link="internal" target="_blank">Chintan Shah</a>, <a href="https://developer.nvidia.com/blog/author/pmedikeri/" title="Posts by Piyush Medikeri" class="author url fn" rel="author follow noopener" data-wpel-link="internal" target="_blank">Piyush Medikeri</a>, <a href="https://developer.nvidia.com/blog/author/avasquez/" title="Posts by Allyson Vasquez" class="author url fn" rel="author follow noopener" data-wpel-link="internal" target="_blank">Allyson Vasquez</a>, <a href="https://developer.nvidia.com/blog/author/tylerz/" title="Posts by Yue Zhu" class="author url fn" rel="author follow noopener" data-wpel-link="internal" target="_blank">Yue Zhu</a>, <a href="https://developer.nvidia.com/blog/author/morganh/" title="Posts by Morgan Huang" class="author url fn" rel="author follow noopener" data-wpel-link="internal" target="_blank">Morgan Huang</a> and <a href="https://developer.nvidia.com/blog/author/bizhao/" title="Posts by Bin Zhao" class="author url fn" rel="author follow noopener" data-wpel-link="internal" target="_blank">Bin Zhao</a> </div> <div class="card--post-attributes-secondary card--post-attributes-header"> <div class="post--rate secondary--attribute"> <div class="wpulike wpulike-heart " ><div class="wp_ulike_general_class wp_ulike_is_not_liked"><button type="button" aria-label="Like Button" data-ulike-id="68713" data-ulike-nonce="4c7ce7a55d" data-ulike-type="post" data-ulike-template="wpulike-heart" data-ulike-display-likers="" data-ulike-likers-style="popover" class="wp_ulike_btn wp_ulike_put_text wp_post_btn_68713"><span><i class="far fa-thumbs-up"></i></span> </button><span class="count-box wp_ulike_counter_up" data-ulike-counter-value="+12"></span> </div></div> <span class="js-like-click like-click"> Like </span> </div> <div class="post--comments-count secondary--attribute"> <a href="#entry-content-comments"> <i class="fad fa-comment-alt-lines"></i> Discuss (0) </a> </div> </div> </div> <img width="1920" height="1080" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="" decoding="async" fetchpriority="high" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail.png 1920w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-625x352.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-196x110.png 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1024x576.png 1024w" sizes="(max-width: 1920px) 100vw, 1920px" /> </div> </div> </div> </div> <div class="col-lg-3 col-md-3 mt-0 mb-0"></div> <div class="main-content col-lg-9 col-md-9 mt-0"> <div class="entry-meta-social"> <ul class="entry-meta-social-links-list"> <li><a data-wpel-link="external" href="https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-linkedin" target="_blank" rel="follow external noopener">L</a></li> <li><a data-wpel-link="external" href="https://twitter.com/intent/tweet?text=Customizing+AI+Models%3A+Train+Character+Detection+and+Recognition+Models+with+NVIDIA+TAO+%7C+NVIDIA+Technical+Blog+https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-twitter" target="_blank" rel="follow external noopener">T</a></li> <li><a data-wpel-link="external" href="https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-facebook" target="_blank" rel="follow external noopener">F</a></li> <li><a data-wpel-link="external" href="https://www.reddit.com/submit?url=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F&title=Customizing+AI+Models%3A+Train+Character+Detection+and+Recognition+Models+with+NVIDIA+TAO+%7C+NVIDIA+Technical+Blog" class="for-reddit" target="_blank" rel="follow external noopener">R</a></li> <li><a href="mailto:?subject=I'd like to share a link with you&body=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-mail">E</a></li> </ul> </div> <div class="entry-content"> <p>Optical Character Detection (OCD) and Optical Character Recognition (OCR) are computer vision techniques used to extract text from images. Use cases vary across industries and include extracting data from scanned documents or forms with handwritten texts, automatically recognizing license plates, sorting boxes or objects in a fulfillment center based on serial numbers, identifying components for inspection on assembly lines based on part numbers, and more. </p> <p>OCR is used in many industries, including financial services, healthcare, logistics, industrial inspection, and smart cities. OCR improves productivity and increases operational efficiency for businesses by automating manual tasks. </p> <p>To be effective, OCR must achieve or exceed human-level accuracy. It is inherently complicated due to the unique use cases it works across. For example, when OCR is analyzing text, the text can vary in font, size, color, shape, and orientation, and can be handwritten or have other noise like partial occlusion. Fine-tuning the model on the test environment becomes extremely important to maintain high accuracy and reduce error rate.  </p> <p><a href="https://developer.nvidia.com/tao-toolkit" data-wpel-link="internal" target="_blank" rel="follow noopener">NVIDIA TAO Toolkit</a> is a low-code AI toolkit that can help developers customize and optimize models for many vision AI applications. NVIDIA introduced new models and features for automating character detection and recognition in TAO 5.0. These models and features will accelerate the creation of custom OCR solutions. For more details, see <a href="https://developer.nvidia.com/blog/access-the-latest-in-vision-ai-model-development-workflows-with-nvidia-tao-toolkit-5-0-2/" data-wpel-link="internal" target="_blank" rel="follow noopener">Access the Latest in Vision AI Model Development Workflows with NVIDIA TAO Toolkit 5.0</a>.</p><div class='code-block code-block-2' style='margin: 40px auto; text-align: center; display: block; clear: both;'> <div style="background: #efefef 0% 0% no-repeat padding-box;margin-top: -8px;margin-bottom: -8px;border-top:1px solid #ccc;border-bottom:1px solid #ccc;"><div style="max-width:1360px;margin:auto;"><div id="metaslider-id-96052" style="width: 100%; margin: 0 auto;" class="ml-slider-3-60-1 metaslider metaslider-flex metaslider-96052 ml-slider ms-theme-default nav-hidden" role="region" aria-roledescription="Slideshow" aria-label="New Slideshow"> <div id="metaslider_container_96052"> <div id="metaslider_96052" class="flexslider"> <ul aria-live="polite" class="slides"> <li style="display: block; width: 100%;" class="slide-96054 ms-image" aria-roledescription="slide" aria-label="slide-96054"><a href="https://nvda.ws/3WDuyIk" target="_blank" data-wpel-link="external" rel="follow external noopener"><img width="1360" height="180" src="https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1.jpg" class="slider-96052 slide-96054" alt="NVIDIA GTC 2025" rel="" title="NVIDIA GTC 2025" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1.jpg 1360w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-300x40.jpg 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-625x83.jpg 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-179x24.jpg 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-768x102.jpg 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-645x85.jpg 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-500x66.jpg 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-160x21.jpg 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-362x48.jpg 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-831x110.jpg 831w, https://developer-blogs.nvidia.com/wp-content/uploads/2025/02/gtc25-banner-reg-now-1360x180-inpost-1-1024x136.jpg 1024w" sizes="(max-width: 1360px) 100vw, 1360px" /></a></li> </ul> </div> </div> </div></div></div></div> <p><em>This post is part of a series on using NVIDIA TAO and pretrained models to create and deploy custom AI models to accurately detect and recognize handwritten texts. This part explains the training and fine-tuning of character detection and recognition models using TAO. <a href="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-2" data-wpel-link="internal" target="_blank" rel="follow noopener">Part 2</a> walks you through the steps to deploy the model using NVIDIA Triton. The steps presented can be used with any other OCR tasks.</em></p> <h2 id="nvidia_tao_ocdocr_workflow" class="wp-block-heading">NVIDIA TAO OCD/OCR workflow<a href="#nvidia_tao_ocdocr_workflow" class="heading-anchor-link"><i class="fas fa-link"></i></a></h2> <figure class="wp-block-image aligncenter size-full"><img decoding="async" width="1183" height="250" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet.png" alt="A workflow overview using OCDNet for generating bounding boxes around areas of text in an image, using the text rectifier to correct any text that is distorted or at extreme angles, then lastly using OCRNet to recognize those sequences of text. " class="wp-image-68721" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet.png 1183w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-300x63.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-625x132.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-179x38.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-768x162.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-645x136.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-500x106.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-160x34.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-362x77.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-521x110.png 521w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/character-recognition-pipeline-ocdnet-ocrnet-1024x216.png 1024w" sizes="(max-width: 1183px) 100vw, 1183px" /><figcaption class="wp-element-caption"><em>Figure 1. Character recognition pipeline with OCDNet and OCRNet</em></figcaption></figure> <p>A pretrained model has been trained on large datasets and can be further fine-tuned with additional data to accomplish a specific task. The Optical Character Detection Network (OCDNet) is a TAO pretrained model that detects text in images with complex backgrounds. It uses a process called differentiable binarization to help accurately locate text of various shapes, sizes, and fonts. The result is a bounding box with the detected text.</p> <p>A text rectifier is middleware that serves as a bridge between character detection and character recognition during the inference phase. Its primary function is to improve the accuracy of recognizing characters on texts that are at extreme angles. To achieve this, the text rectifier takes the vertices of polygons that cover the text area and the original images as inputs. </p> <p>The Optical Character Recognition Network (OCRNet) is another TAO pretrained model that can be used to recognize the characters of text that reside in the detected bounding box regions. This model takes the image as network input and produces a sequence of characters as output.</p> <h3 id="prerequisites" class="wp-block-heading">Prerequisites<a href="#prerequisites" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>To follow along with the tutorial, you will need the following:</p> <ul class="wp-block-list"> <li>An <a href="https://catalog.ngc.nvidia.com/" data-wpel-link="internal" target="_blank" rel="follow noopener">NGC account</a></li> <li>The sample <a href="https://github.com/NVIDIA-AI-IOT/tao_toolkit_recipes/blob/main/tao_ocdr/handwritten/ocdr_handwritten.ipynb" data-wpel-link="external" target="_blank" rel="follow external noopener">Jupyter notebook</a> for training an OCD and OCR model.</li> <li>NVIDIA TAO Toolkit 5.0 (Installation instructions are included in the Jupyter notebooks). For a complete set of dependencies and prerequisites, see the <a href="https://docs.nvidia.com/tao/tao-toolkit/text/tao_toolkit_quick_start_guide.html" data-wpel-link="internal" target="_blank" rel="follow noopener">TAO Toolkit Quick Start Guide</a>. </li> </ul> <h3 id="download_the_dataset" class="wp-block-heading">Download the dataset<a href="#download_the_dataset" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>This tutorial fine-tunes the OCD and OCR model to detect and recognize handwritten letters. It works with the <a href="https://fki.tic.heia-fr.ch/databases/iam-handwriting-database" data-wpel-link="external" target="_blank" rel="follow external noopener">IAM Handwriting Database</a>, a large dataset containing various handwritten English text documents. These text samples will be used to train and test handwritten text recognizers for the OCD and OCR models.</p> <figure class="wp-block-image aligncenter size-full"><img decoding="async" width="144" height="78" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/handwritten-word-iam-dataset.png" alt="The handwritten word ‘have’ from the IAM dataset. " class="wp-image-68722" /><figcaption class="wp-element-caption"><em><em>Figure 2. ‌Handwritten word from the IAM dataset</em></em></figcaption></figure> <p>To gain access to this dataset, register your email address on the <a href="https://fki.tic.heia-fr.ch/register" data-wpel-link="external" target="_blank" rel="follow external noopener">IAM registration page</a>.</p> <p>Once registered, download the following datasets from the <a href="https://fki.tic.heia-fr.ch/databases/download-the-iam-handwriting-database" data-wpel-link="external" target="_blank" rel="follow external noopener">downloads page</a>:</p> <ol class="wp-block-list"> <li>data/ascii.tgz</li> <li>data/formsA-D.tgz</li> <li>data/formsE-H.tgz</li> <li>data/formsI-Z.tgz</li> </ol> <p>The following section explores various aspects of the Jupyter notebook to delve deeper into the fine-tuning process of OCDNet and OCRNet for the purpose of detecting and recognizing handwritten characters.</p> <p>Note that this dataset may be used for noncommercial research purposes only. For more details, review the terms of use on the <a href="https://fki.tic.heia-fr.ch/databases/download-the-iam-handwriting-database" data-wpel-link="external" target="_blank" rel="follow external noopener">IAM Handwriting Database</a>. </p> <h3 id="run_the_notebook" class="wp-block-heading">Run the notebook<a href="#run_the_notebook" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>The OCDR Jupyter notebook showcases how to fine-tune the OCD and OCR models to the IAM handwritten dataset. It also shows how to run inference on the trained models and perform deployment.</p> <h4 class="wp-block-heading">Set up environment variables</h4> <p>Set up the following environment variables in the Jupyter notebook to match your current directory, then execute:</p> <pre class="wp-block-code"><code>%env LOCAL_PROJECT_DIR=home/<username>/ocdr_notebook %env NOTEBOOK_DIR=home/<username>/ocdr_notebook # Set this path if you don't run the notebook from the samples directory. %env NOTEBOOK_ROOT=home/<username>/ocdr_notebook</code></pre> <p>The following folders will be generated:</p> <ul class="wp-block-list"> <li><strong>HOST_DATA_DIR</strong> contains the train/test split data for model training.</li> <li><strong>HOST_SPECS_DIR</strong> houses the specification files that contain the hyperparameters used by TAO to perform training, inference, evaluation, and model deployment.</li> <li><strong>HOST_RESULTS_DIR </strong>contains the results of the fine-tuned OCD and OCR models.</li> <li><strong>PRE_DATA_DIR</strong> is where the downloaded handwritten dataset files will be located. This path will be called to preprocess the data for OCD/OCR model training.</li> </ul> <p>TAO Launcher uses Docker containers when running tasks. For data and results to be visible to Docker, map the location of our local folders to the Docker container using the <code>~/.tao_mounts.json</code> file. Run the cell in the Jupyter notebook to generate the <code>~/.tao_mounts.json</code> file. </p> <p>The environment is now ready for use with the TAO Launcher. The next steps will prepare the handwritten dataset to be in the correct format for TAO OCD model training.</p> <h3 id="prepare_the_dataset_for_ocd_and_ocr" class="wp-block-heading">Prepare the dataset for OCD and OCR<a href="#prepare_the_dataset_for_ocd_and_ocr" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Preprocess the IAM handwritten dataset to match the TAO image format following the steps below. Note that in the folder structure for OCD and OCR model training in TAO,<code> /img</code> houses the handwritten image data, and <code>/gt</code> contains ground truth labels of the characters found in each image. </p> <pre class="wp-block-code"><code>|── train | ├──img | ├──gt |── test | ├──img | ├──gt</code></pre> <p>Begin by moving the four downloaded .tgz files to the location of your <code>$PRE_DATA_DIR</code> directory. If you are following the same steps as above, the .tgz files will be placed in <code>/data/iamdata</code>.</p> <p>Extract the images and ground truth labels from these files. The subsequent cells will extract the image files and move them to the proper folder format when run.</p> <pre class="wp-block-code"><code>!tar -xf $PRE_DATA_DIR/ascii.tgz --directory $PRE_DATA_DIR/ words.txt # Create directories to hold the image data and ground truth files. !mkdir -p $PRE_DATA_DIR/train/img !mkdir -p $PRE_DATA_DIR/test/img !mkdir -p $PRE_DATA_DIR/train/gt !mkdir -p $PRE_DATA_DIR/test/gt </code></pre> <pre class="wp-block-code"><code># Unpack the images, let's use the first two groups of images for training, and the last for validation. !tar -xzf $PRE_DATA_DIR/formsA-D.tgz --directory $PRE_DATA_DIR/train/img !tar -xzf $PRE_DATA_DIR/formsE-H.tgz --directory $PRE_DATA_DIR/train/img !tar -xzf $PRE_DATA_DIR/formsI-Z.tgz --directory $PRE_DATA_DIR/test/img</code></pre> <p>The data is now organized correctly. However, the ground truth label used by IAM dataset is currently in the following format:</p> <pre class="wp-block-code"><code>a01-000u-00-00 ok 154 1 408 768 27 51 AT A # a01-000u-00-00 -> word id for line 00 in form a01-000u # ok -> result of word segmentation # ok: word was correctly # er: segmentation of word can be bad # # 154 -> graylevel to binarize the line containing this word # 1 -> number of components for this word # 408 768 27 51 -> bounding box around this word in x,y,w,h format # AT -> the grammatical tag for this word, see the # file tagset.txt for an explanation # A -> the transcription for this word </code></pre> <p>The <code>words.txt</code> file looks like this:</p> <pre class="wp-block-code"><code> 0 1 0 a01-000u-00-00 ok 154 408 768 27 51 AT A 1 a01-000u-00-01 ok 154 507 766 213 48 NN MOVE 2 a01-000u-00-02 ok 154 796 764 70 50 TO to ... </code></pre> <p>Currently, <code>words.txt</code> uses a four-point coordinate system for drawing a bounding box around the word in an image. TAO requires the use of an eight-point coordinate system to draw a bounding box around detected text. </p> <p>To convert the data to the eight-point coordinate system, use the <code>extract_columns</code> and <code>process_text_file</code> functions provided in section 2.1 of the notebook. <code>words.txt</code> will be transformed into the following DataFrame and will be ready for fine-tuning on an OCDNet model.</p> <pre class="wp-block-code"><code> filename x y x2 y2 x3 y3 x4 y4 word 0 gt_a01-000u.txt 408 768 435 768 435 819 408 819 A 1 gt_a01-000u.txt 507 766 720 766 720 814 507 814 MOVE 2 gt_a01-000u.txt 796 764 866 764 866 814 796 814 to ...</code></pre> <p>To prepare the dataset for OCRNet, the raw image data and labels must be converted to LMDB format, which converts the images and labels into a key-value memory database.</p> <pre class="wp-block-code"><code># Convert the raw train and test dataset to lmdb print("Converting the training set to LMDB.") !tao model ocrnet dataset_convert -e $SPECS_DIR/ocr/experiment.yaml \ dataset_convert.input_img_dir=$DATA_DIR/train/processed \ dataset_convert.gt_file=$DATA_DIR/train/gt.txt \ dataset_convert.results_dir=$DATA_DIR/train/lmdb # Convert the raw test dataset to lmdb print("Converting the testing set to LMDB.") !tao model ocrnet dataset_convert -e $SPECS_DIR/ocr/experiment.yaml \ dataset_convert.input_img_dir=$DATA_DIR/test/processed \ dataset_convert.gt_file=$DATA_DIR/test/gt.txt \ dataset_convert.results_dir=$DATA_DIR/test/lmdb </code></pre> <p>The data is now processed and ready to be fine-tuned on the OCDNet and OCRNet pretrained models.</p> <h2 id="create_a_custom_character_detection_ocd_model" class="wp-block-heading">Create a custom character detection (OCD) model<a href="#create_a_custom_character_detection_ocd_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h2> <p>The NGC CLI will be used to download the pretrained OCDNet model. For more information, visit <a href="http://ngc.nvidia.com" data-wpel-link="internal" target="_blank" rel="follow noopener">NGC</a> and click on Setup in the navigation bar.</p> <h3 id="download_the_ocdnet_pretrained_model" class="wp-block-heading">Download the OCDNet pretrained model<a href="#download_the_ocdnet_pretrained_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <pre class="wp-block-code"><code>!mkdir -p $HOST_RESULTS_DIR/pretrained_ocdnet/ # Pulls pretrained models from NGC !ngc registry model download-version nvidia/tao/ocdnet:trainable_resnet18_v1.0 --dest $HOST_RESULTS_DIR/pretrained_ocdnet/</code></pre> <p>You can check that the model has been downloaded to /pretrained_ocdnet/ using the following call:</p> <pre class="wp-block-code"><code>print("Check that model is downloaded into dir.") !ls -l $HOST_RESULTS_DIR/pretrained_ocdnet/ocdnet_vtrainable_resnet18_v1.0 </code></pre> <h3 id="ocd_training_specification" class="wp-block-heading">OCD training specification<a href="#ocd_training_specification" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>In the specs folder, you can find different files related to how you want to train, evaluate, infer, and export data for both models. For training OCDNet, you will use the train.yaml file in the specs/ocd folder. You can experiment with changing different hyperparameters, such as number of epochs, in this spec file. </p> <p>Below is a code example of some of the configs that you can experiment with:</p> <pre class="wp-block-code"><code>num_gpus: 1 model: load_pruned_graph: False pruned_graph_path: '/results/prune/pruned_0.1.pth' pretrained_model_path: '/data/ocdnet/ocdnet_deformable_resnet18.pth' backbone: deformable_resnet18 train: results_dir: /results/train num_epochs: 300 checkpoint_interval: 1 validation_interval: 1 ... </code></pre> <h3 id="train_the_character_detection_model" class="wp-block-heading">Train the character detection model<a href="#train_the_character_detection_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Now that the specification files are configured, provide the paths to the spec file, the pretrained model, and the results:</p> <pre class="wp-block-code"><code>#Train using TAO Launcher #print("Run training with ngc pretrained model.") !tao model ocdnet train \ -e $SPECS_DIR/train.yaml \ -r $RESULTS_DIR/train \ model.pretrained_model_path=$DATA_DIR/ocdnet_deformable_resnet18.pth</code></pre> <p>Training output will resemble the following. Note that this step could take some time, depending on the number of epochs specified in <code>train.yaml</code>.</p> <pre class="wp-block-code"><code>LOCAL_RANK: 0 - CUDA_VISIBLE_DEVICES: [0] | Name | Type | Params -------------------------------- 0 | model | Model | 12.8 M -------------------------------- 12.8 M Trainable params 0 Non-trainable params 12.8 M Total params 51.106 Total estimated model params size (MB) Training: 0it [00:00, ?it/s]Starting Training Loop. Epoch 0: 100%|█████████| 751/751 [19:57<00:00, 1.59s/it, loss=1.61, v_num=0]</code></pre> <h3 id="evaluate_the_model" class="wp-block-heading">Evaluate the model<a href="#evaluate_the_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Next, evaluate the OCDNet model trained on the IAM dataset.</p> <pre class="wp-block-code"><code># Evaluate on model !tao model ocdnet evaluate \ -e $SPECS_DIR/evaluate.yaml \ evaluate.checkpoint=$RESULTS_DIR/train/model_best.pth</code></pre> <p>Evaluation output will look like the following:</p> <pre class="wp-block-code"><code>test model: 100%|██████████████████████████████| 488/488 [06:44<00:00, 1.21it/s] Precision: 0.9412259824693795 Recall: 0.8738614928590677 Hmean: 0.9062936622138628 Evaluation finished successfully. </code></pre> <h3 id="ocd_inference" class="wp-block-heading">OCD inference<a href="#ocd_inference" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>The inference tool produces annotated image outputs and .txt files that contain prediction information. Run the inference tool below to generate inferences on OCDNet models and visualize the results for detected text.</p> <pre class="wp-block-code"><code># Run inference using TAO !tao model ocdnet inference \ -e $SPECS_DIR/ocd/inference.yaml \ inference.checkpoint=$RESULTS_DIR/ocd/train/model_best.pth \ inference.input_folder=$DATA_DIR/test/img \ inference.results_dir=$RESULTS_DIR/ocd/inference</code></pre> <p>Figure 3 shows the OCDNet inference on a test sample image.</p> <figure class="wp-block-image aligncenter size-full"><img loading="lazy" decoding="async" width="1165" height="814" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output.png" alt="Handwritten text output from OCDNet inference. Bounding boxes are applied to detected words such as ‘discuss’ and ‘best.’ " class="wp-image-69010" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output.png 1165w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-300x210.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-625x437.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-165x115.png 165w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-768x537.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-645x451.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-429x300.png 429w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-129x90.png 129w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-362x253.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-157x110.png 157w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocdnet-inference-output-1024x715.png 1024w" sizes="auto, (max-width: 1165px) 100vw, 1165px" /><figcaption class="wp-element-caption"><em><em>Figure 3. Output from OCDNet inference</em></em></figcaption></figure> <h3 id="export_the_ocd_model_for_deployment" class="wp-block-heading">Export the OCD model for deployment<a href="#export_the_ocd_model_for_deployment" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>The last step is to export the OCD model to ONNX format for deployment.</p> <pre class="wp-block-code"><code>!tao model ocdnet export \ -e $SPECS_DIR/export.yaml \ export.checkpoint=$RESULTS_DIR/train/model_best.pth \ export.onnx_file=$RESULTS_DIR/export/model_best.onnx</code></pre> <h2 id="create_a_custom_character_recognition_ocr_model" class="wp-block-heading">Create a custom character recognition (OCR) model<a href="#create_a_custom_character_recognition_ocr_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h2> <p>Now that you have the trained OCDNet model to detect and apply bounding boxes to areas of handwritten text, use TAO to fine-tune the OCRNet model to recognize and classify the detected letters.</p> <h3 id="download_the_ocrnet_pretrained_model" class="wp-block-heading">Download the OCRNet pretrained model<a href="#download_the_ocrnet_pretrained_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Continuing in the Jupyter notebook, the OCRNet pretrained model will be pulled from NGC CLI.</p> <pre class="wp-block-code"><code>!mkdir -p $HOST_RESULTS_DIR/pretrained_ocrnet/ # Pull pretrained model from NGC !ngc registry model download-version nvidia/tao/ocrnet:trainable_v1.0 --dest $HOST_RESULTS_DIR/pretrained_ocrnet</code></pre> <h3 id="ocr_training_specification" class="wp-block-heading">OCR training specification<a href="#ocr_training_specification" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>OCRNet will use the experiment.yaml spec file to perform training. You can change training hyperparameters such as batch size, number of epochs, and learning rate shown below:</p> <pre class="wp-block-code"><code>dataset: train_dataset_dir: [] val_dataset_dir: /data/test/lmdb character_list_file: /data/character_list max_label_length: 25 batch_size: 32 workers: 4 train: seed: 1111 gpu_ids: [0] optim: name: "adadelta" lr: 0.1 clip_grad_norm: 5.0 num_epochs: 10 checkpoint_interval: 2 validation_interval: 1 </code></pre> <h3 id="train_the_character_recognition_model" class="wp-block-heading">Train the character recognition model<a href="#train_the_character_recognition_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Train the OCRNet model on the dataset. You can also configure spec parameters like the number of epochs or learning rate within the train command, shown below.</p> <pre class="wp-block-code"><code>!tao model ocrnet train -e $SPECS_DIR/ocr/experiment.yaml \ train.results_dir=$RESULTS_DIR/ocr/train \ train.pretrained_model_path=$RESULTS_DIR/pretrained_ocrnet/ocrnet_vtrainable_v1.0/ocrnet_resnet50.pth \ train.num_epochs=20 \ train.optim.lr=1.0 \ dataset.train_dataset_dir=[$DATA_DIR/train/lmdb] \ dataset.val_dataset_dir=$DATA_DIR/test/lmdb \ dataset.character_list_file=$DATA_DIR/train/character_list.txt </code></pre> <p>The output will resemble the following:</p> <pre class="wp-block-code"><code>... Epoch 19: 100%|█| 3605/3605 [08:04<00:00, 7.44it/s, loss=0.0368, v_num=1, val_lCurrent_accuracy : 0.778 Best_accuracy : 0.727 +----------------+--------------+---------------------+ | Ground Truth | Prediction | Confidence && T/F | |----------------+--------------+---------------------| | at | al | 0.2867 False | | home | home | 0.7792 True | | . | . | 0.9828 True | | there | there | 0.5470 True | | had | had | 0.6234 True | +----------------+--------------+---------------------+ </code></pre> <h3 id="evaluate_the_model" class="wp-block-heading">Evaluate the model<a href="#evaluate_the_model" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>You can evaluate the OCRNet model based on the accuracy of its character recognition. Recognition accuracy simply means a percentage of all the characters in a text area that were recognized correctly.</p> <pre class="wp-block-code"><code>!tao model ocrnet evaluate -e $SPECS_DIR/ocr/experiment.yaml \ evaluate.results_dir=$RESULTS_DIR/ocr/evaluate \ evaluate.checkpoint=$RESULTS_DIR/ocr/train/best_accuracy.pth \ evaluate.test_dataset_dir=$DATA_DIR/test/lmdb \ dataset.character_list_file=$DATA_DIR/train/character_list.txt</code></pre> <h3 id="evaluation " class="wp-block-heading">Evaluation<em> </em><a href="#evaluation " class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>The output should appear similar to the following:</p> <pre class="wp-block-code"><code>data directory: /data/iamdata/test/lmdb num samples: 37109 Accuracy: 77.8% </code></pre> <h3 id="ocr_inference" class="wp-block-heading">OCR inference<a href="#ocr_inference" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Inference on OCR will produce a sequence output of recognized characters from the bounding boxes, shown below.</p> <pre class="wp-block-code"><code>!tao model ocrnet inference -e $SPECS_DIR/ocr/experiment.yaml \ inference.results_dir=$RESULTS_DIR/ocr/inference \ inference.checkpoint=$RESULTS_DIR/ocr/train/best_accuracy.pth \ inference.inference_dataset_dir=$DATA_DIR/test/processed \ dataset.character_list_file=$DATA_DIR/train/character_list.txt </code></pre> <pre class="wp-block-code"><code>+--------------------------------------+--------------------+--------------------+ | image_path | predicted_labels | confidence score | |--------------------------------------+--------------------+--------------------| | /data/test/processed/l04-012_28.jpg | lelly | 0.3799 | | /data/test/processed/k04-068_26.jpg | not | 0.9644 | | /data/test/processed/l04-062_58.jpg | set | 0.9542 | | /data/test/processed/l07-176_39.jpg | boat | 0.4693 | | /data/test/processed/k04-039_39.jpg | . | 0.9286 | +--------------------------------------+--------------------+--------------------+ </code></pre> <h3 id="export_ocr_model_for_deployment" class="wp-block-heading">Export OCR model for deployment<a href="#export_ocr_model_for_deployment" class="heading-anchor-link"><i class="fas fa-link"></i></a></h3> <p>Finally, export the OCD Model to ONNX format for deployment.</p> <pre class="wp-block-code"><code>!tao model ocrnet export -e $SPECS_DIR/ocr/experiment.yaml \ export.results_dir=$RESULTS_DIR/ocr/export \ export.checkpoint=$RESULTS_DIR/ocr/train/best_accuracy.pth \ export.onnx_file=$RESULTS_DIR/ocr/export/ocrnet.onnx \ dataset.character_list_file=$DATA_DIR/train/character_list.txt </code></pre> <h2 id="results" class="wp-block-heading">Results<a href="#results" class="heading-anchor-link"><i class="fas fa-link"></i></a></h2> <p>Table 1 highlights the accuracy and performance of the two models featured in this post. The character detection model is fine-tuned on the ICDAR pretrained OCDNet model and character recognition model is fine-tuned on the <a href="https://s3-us-west-2.amazonaws.com/uber-common-public/ubertext/index.html" data-wpel-link="external" target="_blank" rel="follow external noopener">Uber-text</a> OCRNet pretrained model. ICDAR and Uber-text are publicly available datasets that we used to pretrain the OCDNet and OCRNet models, respectively. Both models are available on <a href="https://www.nvidia.com/en-us/gpu-cloud/" data-wpel-link="internal" target="_blank" rel="follow noopener">NGC</a>.  </p> <figure class="wp-block-table aligncenter"><table><tbody><tr><td></td><td><strong>OCDNet</strong></td><td><strong>OCRNet</strong></td></tr><tr><td>Dataset</td><td colspan="2"><a href="https://fki.tic.heia-fr.ch/databases/iam-handwriting-database" data-wpel-link="external" target="_blank" rel="follow external noopener">IAM Handwritten Dataset</a></td></tr><tr><td>Backbone</td><td>Deformable Conv ResNet18</td><td>ResNet50</td></tr><tr><td>Accuracy</td><td>90%</td><td>78%</td></tr><tr><td>Inference resolution</td><td>1024×1024</td><td>1x32x100</td></tr><tr><td>Inference performance (FPS) on NVIDIA L4 GPU</td><td>125 FPS (BS=1)</td><td>8030 (BS=128)</td></tr></tbody></table><figcaption class="wp-element-caption"><em>Table 1. Performance and accuracy data for OCDNet and OCRNet</em></figcaption></figure> <h2 id="summary" class="wp-block-heading">Summary<a href="#summary" class="heading-anchor-link"><i class="fas fa-link"></i></a></h2> <p>This post explains the end-to-end workflow for creating custom character detection and recognition models in NVIDIA TAO. You can start with a pretrained model for character detection (OCDNet) and character recognition (OCRNet) from NGC. Then fine-tune it on your custom dataset using TAO and export the model for inference. </p> <p>Continue reading <a href="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-2" data-wpel-link="internal" target="_blank" rel="follow noopener">Part 2</a> for a step-by-step walkthrough on deploying this model into production using <a href="https://developer.nvidia.com/triton-inference-server" data-wpel-link="internal" target="_blank" rel="follow noopener">NVIDIA Triton</a>.</p>  </div> <div class="block--prospero-assets"> <h2 class="h--smaller txt-clr--blck"> Related resources </h2> <ul><li>DLI course: <a class='wpel-ignore' target='_blank' href='https://www.nvidia.com/en-us/training/instructor-led-workshops/building-conversational-ai-apps/?ncid=em-nurt-245273-vt33'>Building Conversational AI Applications </a></li><li>GTC session: <a class='wpel-ignore' target='_blank' href='https://www.nvidia.com/gtc/session-catalog/?tab.catalogallsessionstab=1700692987788001F1cG&search=S72653&ncid=em-even-124008-vt33-23spring#/'>Trains as a Sensor: Real-time Vision AI system in Ruggedized Environments</a></li><li>GTC session: <a class='wpel-ignore' target='_blank' href='https://www.nvidia.com/gtc/session-catalog/?tab.catalogallsessionstab=1700692987788001F1cG&search=DLIW73640&ncid=em-even-124008-vt33-23spring#/'>Efficient Large Language Model Customization</a></li><li>GTC session: <a class='wpel-ignore' target='_blank' href='https://www.nvidia.com/gtc/session-catalog/?tab.catalogallsessionstab=1700692987788001F1cG&search=S71296&ncid=em-even-124008-vt33-23spring#/'>Leverage Tailored LLMs for Improved Insights in Rail Transportation</a></li><li>SDK: <a class='wpel-ignore' target='_blank' href='https://developer.nvidia.com/tao-toolkit?ncid=em-nurt-245273-vt33'>TAO Toolkit</a></li><li>Webinar: <a class='wpel-ignore' target='_blank' href='https://gateway.on24.com/wcc/eh/1407606/lp/3798294/?embedUrl=https://www.nvidia.com/en-us/about-nvidia/webinar-portal/'>Enabling Impact at Scale with Conversational AI and Computer Vision</a></li></ul> </div> <div class="card--post-attributes-secondary"> <div class="post--comments-count secondary--attribute"> <a href="#entry-content-comments"> <i class="fad fa-comment-alt-lines"></i> Discuss (0) </a> </div> <div class="post--rate secondary--attribute"> <div class="wpulike wpulike-heart " ><div class="wp_ulike_general_class wp_ulike_is_not_liked"><button type="button" aria-label="Like Button" data-ulike-id="68713" data-ulike-nonce="4c7ce7a55d" data-ulike-type="post" data-ulike-template="wpulike-heart" data-ulike-display-likers="" data-ulike-likers-style="popover" class="wp_ulike_btn wp_ulike_put_text wp_post_btn_68713"><span><i class="far fa-thumbs-up"></i></span> </button><span class="count-box wp_ulike_counter_up" data-ulike-counter-value="+12"></span> </div></div> <span class="js-like-click like-click"> Like </span> </div> </div> <div class="card--post-attributes-secondary tags"> <div class="caption"> <h2 class="h--smaller txt-clr--blck">Tags</h2> </div> <div class="content-s post-tags--list mt-0"> <a href="https://developer.nvidia.com/blog/category/computer-vision/" data-wpel-link="internal" target="_blank" rel="follow noopener">Computer Vision / Video Analytics</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?industry=Retail+%2F+Consumer+Packaged+Goods" data-wpel-link="internal" target="_blank" rel="follow noopener">Retail / Consumer Packaged Goods</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?products=Metropolis" data-wpel-link="internal" target="_blank" rel="follow noopener">Metropolis</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?products=TAO+Toolkit" data-wpel-link="internal" target="_blank" rel="follow noopener">TAO Toolkit</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?products=Triton+Inference+Server" data-wpel-link="internal" target="_blank" rel="follow noopener">Triton Inference Server</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?learning_levels=Intermediate+Technical" data-wpel-link="internal" target="_blank" rel="follow noopener">Intermediate Technical</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?content_types=Deep+dive" data-wpel-link="internal" target="_blank" rel="follow noopener">Deep dive</a> | <a href="https://developer.nvidia.com/blog/recent-posts/?content_types=Tutorial" data-wpel-link="internal" target="_blank" rel="follow noopener">Tutorial</a> | <a href="https://developer.nvidia.com/blog/tag/deep-learning/" data-wpel-link="internal" target="_blank" rel="follow noopener">deep learning</a> | <a href="https://developer.nvidia.com/blog/tag/featured/" data-wpel-link="internal" target="_blank" rel="follow noopener">featured</a> | <a href="https://developer.nvidia.com/blog/tag/pre-trained-foundation-models/" data-wpel-link="internal" target="_blank" rel="follow noopener">Pre-Trained / Foundation Models</a> </div> </div> <div class="post-authors-list"> <div class="entry-content-author"> <div class="caption"> <h2 class="h--smaller txt-clr--blck"> About the Authors </h2> </div> <div class="media author-info"> <div class="author-media-left media-left"> <img alt='Avatar photo' src='https://developer-blogs.nvidia.com/wp-content/uploads/2020/01/cropped-ChintanShah-131x131.jpg' srcset='https://developer-blogs.nvidia.com/wp-content/uploads/2020/01/cropped-ChintanShah-262x262.jpg 2x' class='avatar avatar-131 photo' height='131' width='131' loading='lazy' decoding='async'/> </div> <div class="author-media-body media-body"> <b> About Chintan Shah </b> <br/> Chintan Shah is a senior product manager at NVIDIA, focusing on AI products. He manages NVIDIA TAO and vision AI solutions for smart spaces, retail, industrial, and other industry verticals. His focus is on simplifying and democratizing AI for all developers and enterprises. <div id="author-link"> <a href="https://developer.nvidia.com/blog/author/chintan-shah/" rel="author follow noopener" data-wpel-link="internal" target="_blank"> View all posts by Chintan Shah<svg width="16" height="16" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 612"><path d="M305 239c9.4 9.4 9.4 24.6 0 33.9L113 465c-9.4 9.4-24.6 9.4-33.9 0s-9.4-24.6 0-33.9l175-175L79 81c-9.4-9.4-9.4-24.6 0-33.9s24.6-9.4 33.9 0L305 239z"/></svg> </a> </div> </div> </div> <div class="media author-info"> <div class="author-media-left media-left"> <img alt='Piyush Medikeri' src='https://developer-blogs.nvidia.com/wp-content/uploads/2023/07/cropped-piyush-medikeri-131x131.jpg' srcset='https://developer-blogs.nvidia.com/wp-content/uploads/2023/07/cropped-piyush-medikeri-262x262.jpg 2x' class='avatar avatar-131 photo' height='131' width='131' loading='lazy' decoding='async'/> </div> <div class="author-media-body media-body"> <b> About Piyush Medikeri </b> <br/> Piyush Medikeri is a senior systems software engineer at NVIDIA focused on improving developer experiences on various NVIDIA AI software solutions in the areas of video analytics, robotics, simulation, deep learning, and autonomous vehicles. Before joining NVIDIA, Piyush interned at PLEN Robotics in Osaka and MathWorks, MA as a Robotics Engineer. He holds an master’s degree in Robotics and Autonomous Systems (specialization in AI) from Arizona State University. <div id="author-link"> <a href="https://developer.nvidia.com/blog/author/pmedikeri/" rel="author follow noopener" data-wpel-link="internal" target="_blank"> View all posts by Piyush Medikeri<svg width="16" height="16" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 612"><path d="M305 239c9.4 9.4 9.4 24.6 0 33.9L113 465c-9.4 9.4-24.6 9.4-33.9 0s-9.4-24.6 0-33.9l175-175L79 81c-9.4-9.4-9.4-24.6 0-33.9s24.6-9.4 33.9 0L305 239z"/></svg> </a> </div> </div> </div> <div class="media author-info"> <div class="author-media-left media-left"> <img alt='Allyson Vasquez' src='https://developer-blogs.nvidia.com/wp-content/uploads/2023/07/cropped-allyson-vasquez-131x131.jpg' srcset='https://developer-blogs.nvidia.com/wp-content/uploads/2023/07/cropped-allyson-vasquez-262x262.jpg 2x' class='avatar avatar-131 photo' height='131' width='131' loading='lazy' decoding='async'/> </div> <div class="author-media-body media-body"> <b> About Allyson Vasquez </b> <br/> Allyson Vasquez is a systems software engineer at NVIDIA, responsible for improving the developer experience in the domains of autonomous vehicles, simulation, computer vision, data science, and conversational AI. She holds a master’s degree in Computer Science from the University of North Carolina at Charlotte, specializing in data science. <div id="author-link"> <a href="https://developer.nvidia.com/blog/author/avasquez/" rel="author follow noopener" data-wpel-link="internal" target="_blank"> View all posts by Allyson Vasquez<svg width="16" height="16" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 612"><path d="M305 239c9.4 9.4 9.4 24.6 0 33.9L113 465c-9.4 9.4-24.6 9.4-33.9 0s-9.4-24.6 0-33.9l175-175L79 81c-9.4-9.4-9.4-24.6 0-33.9s24.6-9.4 33.9 0L305 239z"/></svg> </a> </div> </div> </div> <div class="media author-info"> <div class="author-media-left media-left"> <img alt='Avatar photo' src='https://developer-blogs.nvidia.com/wp-content/uploads/2021/02/yue_zhu_tyler-131x131.jpg' srcset='https://developer-blogs.nvidia.com/wp-content/uploads/2021/02/yue_zhu_tyler-262x262.jpg 2x' class='avatar avatar-131 photo' height='131' width='131' loading='lazy' decoding='async'/> </div> <div class="author-media-body media-body"> <b> About Yue Zhu </b> <br/> Yue Zhu is a senior system software engineer at NVIDIA, focusing on perception algorithms development for automotive driving, intelligent video analytics, and robotics. He received a B.S and M.S in computer science from University of Electronic Science and Technology of China. <div id="author-link"> <a href="https://developer.nvidia.com/blog/author/tylerz/" rel="author follow noopener" data-wpel-link="internal" target="_blank"> View all posts by Yue Zhu<svg width="16" height="16" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 612"><path d="M305 239c9.4 9.4 9.4 24.6 0 33.9L113 465c-9.4 9.4-24.6 9.4-33.9 0s-9.4-24.6 0-33.9l175-175L79 81c-9.4-9.4-9.4-24.6 0-33.9s24.6-9.4 33.9 0L305 239z"/></svg> </a> </div> </div> </div> <div class="media author-info"> <div class="author-media-left media-left"> <img alt='Avatar photo' src='https://developer-blogs.nvidia.com/wp-content/uploads/2021/02/Morgan_Huang-131x131.jpg' srcset='https://developer-blogs.nvidia.com/wp-content/uploads/2021/02/Morgan_Huang-262x262.jpg 2x' class='avatar avatar-131 photo' height='131' width='131' loading='lazy' decoding='async'/> </div> <div class="author-media-body media-body"> <b> About Morgan Huang </b> <br/> Morgan Huang is a senior software testing engineer at NVIDIA, focusing on accuracy or performance issue analysis and optimization. He oversees the NVIDIA TLT forum.He holds a master’s degree in electrical engineering from Beijing Jiaotong University, China. <div id="author-link"> <a href="https://developer.nvidia.com/blog/author/morganh/" rel="author follow noopener" data-wpel-link="internal" target="_blank"> View all posts by Morgan Huang<svg width="16" height="16" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 612"><path d="M305 239c9.4 9.4 9.4 24.6 0 33.9L113 465c-9.4 9.4-24.6 9.4-33.9 0s-9.4-24.6 0-33.9l175-175L79 81c-9.4-9.4-9.4-24.6 0-33.9s24.6-9.4 33.9 0L305 239z"/></svg> </a> </div> </div> </div> <div class="media author-info"> <div class="author-media-left media-left"> <img alt='Bin Zhao' src='https://developer-blogs.nvidia.com/wp-content/uploads/2023/07/bin-zhao-131x131.jpg' srcset='https://developer-blogs.nvidia.com/wp-content/uploads/2023/07/bin-zhao-262x262.jpg 2x' class='avatar avatar-131 photo' height='131' width='131' loading='lazy' decoding='async'/> </div> <div class="author-media-body media-body"> <b> About Bin Zhao </b> <br/> Bin Zhao is a senior system software engineer at NVIDIA, focusing on perception algorithms development for automotive driving, intelligent video analytics, and robotics. He holds a master’s degree in Mechanical Engineering from Xi’an Jiaotong University, China. <div id="author-link"> <a href="https://developer.nvidia.com/blog/author/bizhao/" rel="author follow noopener" data-wpel-link="internal" target="_blank"> View all posts by Bin Zhao<svg width="16" height="16" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 320 612"><path d="M305 239c9.4 9.4 9.4 24.6 0 33.9L113 465c-9.4 9.4-24.6 9.4-33.9 0s-9.4-24.6 0-33.9l175-175L79 81c-9.4-9.4-9.4-24.6 0-33.9s24.6-9.4 33.9 0L305 239z"/></svg> </a> </div> </div> </div> </div> </div> <div class="entry-content-comments" id="entry-content-comments"> <div class="container" id="pf-disqus-thread"> <div class="row" id="respond"> <div class="col-md-12 related-posts-comments mb-0"><h2 class="h--smaller txt-clr--blck mb-0">Comments</h2></div> </div> <div class="row"> <div class="col-lg-12 col-md-12 col-sm-12 col-xs-12"> <div class="wpdc-comments-loading" id="wpdc-comments" data-post-id="68713"></div> </div> </div> </div> </div> <div id="main-content-end"></div> </div> <div id="sidebar" class="sidebar col-lg-3 col-md-3 mt-0"> <div class="sticky-top-pre"></div> <div class="sidebar-item not-sticky-top"> <div id="sidebar-content" class="sidebar__inner"> <script> window.is_post_page = true; window.nv_homepage_url = 'https://developer.nvidia.com/blog'; </script> <div class="block--related-posts prspr-related" data-source="nv"> <h2 class="h--smaller txt-clr--blck">Related posts</h2> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="74323" data-post-title="Robust Scene Text Detection and Recognition: Implementation"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1999" height="1125" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="Decorative image of a workflow and the text "Part 2"." decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2.png 1999w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-625x352.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-195x110.png 195w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part2-1024x576.png 1024w" sizes="auto, (max-width: 1999px) 100vw, 1999px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Robust Scene Text Detection and Recognition: Implementation </h3> </div> </div> <a href="https://developer.nvidia.com/blog/robust-scene-text-detection-and-recognition-implementation/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Robust Scene Text Detection and Recognition: Implementation</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="74322" data-post-title="Robust Scene Text Detection and Recognition: Introduction"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1999" height="1125" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="Decorative image of a workflow and the text "Part 1"." decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1.png 1999w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-625x352.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-195x110.png 195w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/wipro-series-featured-part1-1024x576.png 1024w" sizes="auto, (max-width: 1999px) 100vw, 1999px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Robust Scene Text Detection and Recognition: Introduction </h3> </div> </div> <a href="https://developer.nvidia.com/blog/robust-scene-text-detection-and-recognition-introduction/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Robust Scene Text Detection and Recognition: Introduction</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="74846" data-post-title="Develop and Optimize Vision AI Models for Trillions of Devices with NVIDIA TAO"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1552" height="873" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1.jpg" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1.jpg 1552w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-300x169.jpg 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-625x352.jpg 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-179x101.jpg 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-768x432.jpg 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-1536x864.jpg 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-645x363.jpg 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-960x540.jpg 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-500x281.jpg 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-160x90.jpg 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-362x204.jpg 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-196x110.jpg 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/12/data-train-deploy-nvidia-tao-graphic-1-1024x576.jpg 1024w" sizes="auto, (max-width: 1552px) 100vw, 1552px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Develop and Optimize Vision AI Models for Trillions of Devices with NVIDIA TAO </h3> </div> </div> <a href="https://developer.nvidia.com/blog/develop-and-optimize-vision-ai-models-for-trillions-of-devices-with-nvidia-tao/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Develop and Optimize Vision AI Models for Trillions of Devices with NVIDIA TAO</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="69017" data-post-title="Customizing AI Models: Deploy a Character Detection and Recognition Model with NVIDIA Triton"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1920" height="1080" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1.png 1920w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-625x352.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-196x110.png 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/08/ocd-ocr-detail-1-1024x576.png 1024w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Customizing AI Models: Deploy a Character Detection and Recognition Model with NVIDIA Triton </h3> </div> </div> <a href="https://developer.nvidia.com/blog/create-custom-character-detection-and-recognition-models-with-nvidia-tao-part-2/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Customizing AI Models: Deploy a Character Detection and Recognition Model with NVIDIA Triton</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="62089" data-post-title="Access the Latest in Vision AI Model Development Workflows with NVIDIA TAO Toolkit 5.0"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1920" height="1080" src="https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic.jpg" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="TAO Toolkit graphic" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic.jpg 1920w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-300x169.jpg 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-625x352.jpg 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-179x101.jpg 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-768x432.jpg 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-1536x864.jpg 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-500x281.jpg 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-160x90.jpg 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-362x204.jpg 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-196x110.jpg 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/01/tao-toolkit-5-graphic-1024x576.jpg 1024w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Access the Latest in Vision AI Model Development Workflows with NVIDIA TAO Toolkit 5.0 </h3> </div> </div> <a href="https://developer.nvidia.com/blog/access-the-latest-in-vision-ai-model-development-workflows-with-nvidia-tao-toolkit-5-0-2/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Access the Latest in Vision AI Model Development Workflows with NVIDIA TAO Toolkit 5.0</span> </a> </div> </div> </div> <div class="block--related-posts hide" data-source="wp"> <h2 class="h--smaller txt-clr--blck">Related posts</h2> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="92268" data-post-title="Processing High-Quality Vietnamese Language Data with NVIDIA NeMo Curator"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1999" height="1124" src="https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="The process of data curation for LLMs." decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured.png 1999w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-625x351.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-196x110.png 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/11/llm-data-curator-featured-1024x576.png 1024w" sizes="auto, (max-width: 1999px) 100vw, 1999px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Processing High-Quality Vietnamese Language Data with NVIDIA NeMo Curator </h3> </div> </div> <a href="https://developer.nvidia.com/blog/processing-high-quality-vietnamese-language-data-with-nvidia-nemo-curator/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Processing High-Quality Vietnamese Language Data with NVIDIA NeMo Curator</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="86238" data-post-title="Access to NVIDIA NIM Now Available Free to Developer Program Members"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1920" height="1080" src="https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="An image showing NVIDIA NIM." decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured.png 1920w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-625x352.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-196x110.png 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/07/nim-visual-featured-1024x576.png 1024w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Access to NVIDIA NIM Now Available Free to Developer Program Members </h3> </div> </div> <a href="https://developer.nvidia.com/blog/access-to-nvidia-nim-now-available-free-to-developer-program-members/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Access to NVIDIA NIM Now Available Free to Developer Program Members</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="81524" data-post-title="Revolutionizing Graph Analytics: Next-Gen Architecture with NVIDIA cuGraph Acceleration"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1920" height="1080" src="https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1.jpg" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1.jpg 1920w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-300x169.jpg 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-625x352.jpg 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-179x101.jpg 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-768x432.jpg 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-1536x864.jpg 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-645x363.jpg 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-960x540.jpg 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-500x281.jpg 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-160x90.jpg 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-362x204.jpg 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-196x110.jpg 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2024/04/rapids-press-wholegraph-blog-gtc24-1920x1080-1-1024x576.jpg 1024w" sizes="auto, (max-width: 1920px) 100vw, 1920px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Revolutionizing Graph Analytics: Next-Gen Architecture with NVIDIA cuGraph Acceleration </h3> </div> </div> <a href="https://developer.nvidia.com/blog/revolutionizing-graph-analytics-next-gen-architecture-with-nvidia-cugraph-acceleration/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Revolutionizing Graph Analytics: Next-Gen Architecture with NVIDIA cuGraph Acceleration</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="71925" data-post-title="Efficient CUDA Debugging: Memory Initialization and Thread Synchronization with NVIDIA Compute Sanitizer"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1821" height="1024" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda.png 1821w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-625x351.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-1536x864.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-500x281.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-196x110.png 196w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/10/debugging-cuda-1024x576.png 1024w" sizes="auto, (max-width: 1821px) 100vw, 1821px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Efficient CUDA Debugging: Memory Initialization and Thread Synchronization with NVIDIA Compute Sanitizer </h3> </div> </div> <a href="https://developer.nvidia.com/blog/efficient-cuda-debugging-memory-initialization-and-thread-synchronization-with-nvidia-compute-sanitizer/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Efficient CUDA Debugging: Memory Initialization and Thread Synchronization with NVIDIA Compute Sanitizer</span> </a> </div> </div> <div class="carousel-row__slide js-post-card related-post--row" data-post-id="71113" data-post-title="Analyzing the Security of Machine Learning Research Code"> <div class="carousel-row-slide__inner"> <div class="carousel-row-slide__thumbnail"> <img width="1634" height="920" src="https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop.png" class="attachment-post-thumbnail size-post-thumbnail wp-post-image" alt="" decoding="async" loading="lazy" srcset="https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop.png 1634w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-300x169.png 300w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-625x352.png 625w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-179x101.png 179w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-768x432.png 768w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-1536x865.png 1536w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-645x363.png 645w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-960x540.png 960w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-500x282.png 500w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-160x90.png 160w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-362x204.png 362w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-195x110.png 195w, https://developer-blogs.nvidia.com/wp-content/uploads/2023/09/man-with-laptop-1024x577.png 1024w" sizes="auto, (max-width: 1634px) 100vw, 1634px" /> </div> <div class="carousel-row-slide__content"> <div class="carousel-row-slide__title"> <h3 class="h--smallest carousel-row-slide__heading"> Analyzing the Security of Machine Learning Research Code </h3> </div> </div> <a href="https://developer.nvidia.com/blog/analyzing-the-security-of-machine-learning-research-code/" class="carousel-row-slide__link" data-wpel-link="internal" target="_blank" rel="follow noopener"> <span class="visually-hidden">Analyzing the Security of Machine Learning Research Code</span> </a> </div> </div> </div> </div> </div> </div> </div> </div> </section>  <script type="text/javascript"> function $(id) {return !id ? null : document.getElementById(id);} loadPrettifyCss = function () { if(!$('prettify_css')) { css = document.createElement('link'); css.id = 'prettify_css'; css.type = 'text/css'; css.rel = 'stylesheet'; css.href = 'https://developer-blogs.nvidia.com/wp-content/plugins/wp-code-prettify/css/prettify.css'; var headNode = document.getElementsByTagName("head")[0]; headNode.appendChild(css); } else { $('prettify_css').href = 'https://developer-blogs.nvidia.com/wp-content/plugins/wp-code-prettify/css/prettify.css'; } if(!$('prettify_custom')) { css = document.createElement('style'); css.id = 'prettify_custom'; css.type = 'text/css'; css.rel = 'stylesheet'; css.innerHTML = 'pre.prettyprint { max-height: 500px; white-space: pre; word-wrap: normal; overflow: auto; background-color: #f8f8f8;border-radius: 3px; } pre.prettyprint[title]:before { float: right; border: 1px solid #ccc; border-radius: 3px; background: #ddd; padding: 10px; content: attr(title); font: 13px "Trebuchet MS", "Helvetica Neue", Helvetica, Arial, sans-serif; }'; var headNode = document.getElementsByTagName("head")[0]; headNode.appendChild(css); } else { $('prettify_css').innerHTML = 'pre.prettyprint { max-height: 500px; white-space: pre; word-wrap: normal; overflow: auto; background-color: #f8f8f8;border-radius: 3px; } pre.prettyprint[title]:before { float: right; border: 1px solid #ccc; border-radius: 3px; background: #ddd; padding: 10px; content: attr(title); font: 13px "Trebuchet MS", "Helvetica Neue", Helvetica, Arial, sans-serif; }'; } } </script> <script type="text/javascript"> loadPrettifyCss(); </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/wp-code-prettify/js/prettify.js"></script> <script type="text/javascript"> function wpCodePrettifyOnLoad(func){ var wpCodePrettifyOldOnLoad = window.onload; if (typeof window.onload != 'function') { window.onload = func } else { window.onload = function () { wpCodePrettifyOldOnLoad(); func() } } } wpCodePrettifyOnLoad(function(){prettyPrint();}); </script>  <footer id="nv-footer"></footer> <script type="text/javascript"> (function() { const defaultLinks = JSON.parse("{\n \"columns\": [],\n \"links\": [\n {\n \"text\": \"Legal Information\",\n \"href\": \"https:\/\/www.nvidia.com\/en-us\/about-nvidia\/legal-info\/\"\n },\n {\n \"text\": \"Terms of Use\",\n \"href\": \"https:\/\/developer.nvidia.com\/legal\/terms\"\n },\n {\n \"text\": \"Privacy Policy\",\n \"href\": \"https:\/\/www.nvidia.com\/en-us\/about-nvidia\/privacy-policy\/\"\n },\n {\n \"text\": \"Cookie Policy\",\n \"href\": \"https:\/\/www.nvidia.com\/en-us\/about-nvidia\/cookie-policy\/\"\n },\n {\n \"text\": \"Contact\",\n \"href\": \"https:\/\/developer.nvidia.com\/contact\"\n }\n ],\n \"news\": {\n \"title\": \"Sign up for NVIDIA News\",\n \"href\": \"https:\/\/developer.nvidia.com\/email-signup\"\n },\n \"social\": {\n \"title\": \"Follow NVIDIA Developer\",\n \"links\": [\n {\n \"text\": \"Facebook\",\n \"href\": \"https:\/\/facebook.com\/nvidiaai\",\n \"provider\": \"facebook\"\n },\n {\n \"text\": \"Instagram\",\n \"href\": \"https:\/\/instagram.com\/nvidiadeveloper\",\n \"provider\": \"instagram\"\n },\n {\n \"text\": \"LinkedIn\",\n \"href\": \"https:\/\/www.linkedin.com\/showcase\/nvidia-ai\",\n \"provider\": \"linkedin\"\n },\n {\n \"text\": \"Twitter\",\n \"href\": \"https:\/\/twitter.com\/nvidiadeveloper\",\n \"provider\": \"twitter\"\n },\n {\n \"text\": \"YouTube\",\n \"href\": \"https:\/\/www.youtube.com\/user\/NVIDIADeveloper\",\n \"provider\": \"youtube\"\n }\n ]\n }\n }\n ") const mountFooterComponent = (locale, response) => { new NVDeveloperFooter({ target: document.getElementById('nv-footer'), props: { locale: locale, menu: response } }) } const handleError = (error, locale, menu) => { console.debug(error) mountFooterComponent(locale, menu) } const initFooter = () => { let menuUrl = 'https://d29g4g2dyqv443.cloudfront.net/menu/en-US/footer.json' let locale = 'en-US'; if (!menuUrl) { handleError('Footer menu URL not defined.', locale, defaultLinks) return } fetch(menuUrl) .then((response) => response.json()) .then((response) => mountFooterComponent(locale, response)) .catch((error) => handleError(error, locale, defaultLinks)) } document.addEventListener('DOMContentLoaded', initFooter); })(); </script> <script type="text/javascript">_satellite.pageBottom();</script>   <script type="text/javascript" src="https://images.nvidia.com/aem-dam/Solutions/ot-js/ot-custom.js"></script>  <div style="display:none"> <div class="grofile-hash-map-47d7ab2b5f8511c5404b77f141a91502"> </div> <div class="grofile-hash-map-2bd80bada414ed381e36dc1642846a0b"> </div> <div class="grofile-hash-map-91b06285f36f913cafd42ed1b2e52fa7"> </div> <div class="grofile-hash-map-1019b385ff67e934a82d8cbec62d4dfe"> </div> <div class="grofile-hash-map-b9afea89de77cfa4afbe6456b77194e8"> </div> <div class="grofile-hash-map-da1b38eb001a9e14acafada818706fd5"> </div> </div> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shCore.js?ver=3.0.9b" id="syntaxhighlighter-core-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushAS3.js?ver=3.0.9b" id="syntaxhighlighter-brush-as3-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushArduino.js?ver=3.0.9b" id="syntaxhighlighter-brush-arduino-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushBash.js?ver=3.0.9b" id="syntaxhighlighter-brush-bash-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushColdFusion.js?ver=3.0.9b" id="syntaxhighlighter-brush-coldfusion-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/third-party-brushes/shBrushClojure.js?ver=20090602" id="syntaxhighlighter-brush-clojure-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushCpp.js?ver=3.0.9b" id="syntaxhighlighter-brush-cpp-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushCSharp.js?ver=3.0.9b" id="syntaxhighlighter-brush-csharp-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushCss.js?ver=3.0.9b" id="syntaxhighlighter-brush-css-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushDelphi.js?ver=3.0.9b" id="syntaxhighlighter-brush-delphi-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushDiff.js?ver=3.0.9b" id="syntaxhighlighter-brush-diff-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushErlang.js?ver=3.0.9b" id="syntaxhighlighter-brush-erlang-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/third-party-brushes/shBrushFSharp.js?ver=20091003" id="syntaxhighlighter-brush-fsharp-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushGo.js?ver=3.0.9b" id="syntaxhighlighter-brush-go-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushGroovy.js?ver=3.0.9b" id="syntaxhighlighter-brush-groovy-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushHaskell.js?ver=3.0.9b" id="syntaxhighlighter-brush-haskell-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushJava.js?ver=3.0.9b" id="syntaxhighlighter-brush-java-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushJavaFX.js?ver=3.0.9b" id="syntaxhighlighter-brush-javafx-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushJScript.js?ver=3.0.9b" id="syntaxhighlighter-brush-jscript-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/third-party-brushes/shBrushLatex.js?ver=20090613" id="syntaxhighlighter-brush-latex-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/third-party-brushes/shBrushMatlabKey.js?ver=20091209" id="syntaxhighlighter-brush-matlabkey-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/third-party-brushes/shBrushObjC.js?ver=20091207" id="syntaxhighlighter-brush-objc-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushPerl.js?ver=3.0.9b" id="syntaxhighlighter-brush-perl-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushPhp.js?ver=3.0.9b" id="syntaxhighlighter-brush-php-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushPlain.js?ver=3.0.9b" id="syntaxhighlighter-brush-plain-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushPowerShell.js?ver=3.0.9b" id="syntaxhighlighter-brush-powershell-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushPython.js?ver=3.0.9b" id="syntaxhighlighter-brush-python-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/third-party-brushes/shBrushR.js?ver=20100919" id="syntaxhighlighter-brush-r-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushRuby.js?ver=3.0.9b" id="syntaxhighlighter-brush-ruby-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushScala.js?ver=3.0.9b" id="syntaxhighlighter-brush-scala-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushSql.js?ver=3.0.9b" id="syntaxhighlighter-brush-sql-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushSwift.js?ver=3.0.9b" id="syntaxhighlighter-brush-swift-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushVb.js?ver=3.0.9b" id="syntaxhighlighter-brush-vb-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushXml.js?ver=3.0.9b" id="syntaxhighlighter-brush-xml-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/scripts/shBrushYaml.js?ver=3.0.9b" id="syntaxhighlighter-brush-yaml-js"></script> <script type='text/javascript'> (function(){ var corecss = document.createElement('link'); var themecss = document.createElement('link'); var corecssurl = "https://developer-blogs.nvidia.com/wp-content/plugins/syntaxhighlighter/syntaxhighlighter3/styles/shCore.css?ver=3.0.9b"; if ( corecss.setAttribute ) { corecss.setAttribute( "rel", "stylesheet" ); corecss.setAttribute( "type", "text/css" ); corecss.setAttribute( "href", corecssurl ); } else { corecss.rel = "stylesheet"; corecss.href = corecssurl; } document.head.appendChild( corecss ); var themecssurl = "https://developer-blogs.nvidia.com/wp-content/plugins/nv-tweaks/css/shThemeNvidia.css?ver=3.0.9b"; if ( themecss.setAttribute ) { themecss.setAttribute( "rel", "stylesheet" ); themecss.setAttribute( "type", "text/css" ); themecss.setAttribute( "href", themecssurl ); } else { themecss.rel = "stylesheet"; themecss.href = themecssurl; } document.head.appendChild( themecss ); })(); SyntaxHighlighter.config.strings.expandSource = '+ expand source'; SyntaxHighlighter.config.strings.help = '?'; SyntaxHighlighter.config.strings.alert = 'SyntaxHighlighter\n\n'; SyntaxHighlighter.config.strings.noBrush = 'Can\'t find brush for: '; SyntaxHighlighter.config.strings.brushNotHtmlScript = 'Brush wasn\'t configured for html-script option: '; SyntaxHighlighter.defaults['gutter'] = false; SyntaxHighlighter.defaults['pad-line-numbers'] = false; SyntaxHighlighter.defaults['toolbar'] = false; SyntaxHighlighter.all(); // Infinite scroll support if ( typeof( jQuery ) !== 'undefined' ) { jQuery( function( $ ) { $( document.body ).on( 'post-load', function() { SyntaxHighlighter.highlight(); } ); } ); } </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-includes/js/jquery/jquery.min.js?ver=3.7.1" id="jquery-core-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-includes/js/jquery/jquery-migrate.min.js?ver=3.4.1" id="jquery-migrate-js"></script> <script type="text/javascript" id="load_comments_js-js-extra"> /* <![CDATA[ */ var wpdc = {"commentsURL":"https:\/\/developer.nvidia.com\/blog\/wp-json\/wp-discourse\/v1\/discourse-comments"}; /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/wp-discourse/lib/../js/load-comments.js?ver=1740539867" id="load_comments_js-js"></script> <script type="text/javascript" id="wp_ulike-js-extra"> /* <![CDATA[ */ var wp_ulike_params = {"ajax_url":"https:\/\/developer-blogs.nvidia.com\/wp-admin\/admin-ajax.php","notifications":"0"}; var wp_ulike_params = {"ajax_url":"https:\/\/developer.nvidia.com\/blog\/wp-admin\/admin-ajax.php"}; /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/wp-ulike/assets/js/wp-ulike.min.js?ver=4.7.7" id="wp_ulike-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/nv-language-switcher/js/nv-language-switcher.js" id="nv-language-switcher-js"></script> <script type="text/javascript" src="https://secure.gravatar.com/js/gprofiles.js?ver=202509" id="grofiles-cards-js"></script> <script type="text/javascript" id="wpgroho-js-extra"> /* <![CDATA[ */ var WPGroHo = {"my_hash":""}; /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/jetpack/modules/wpgroho.js?ver=14.0" id="wpgroho-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/page-links-to/dist/new-tab.js?ver=3.3.5" id="page-links-to-js"></script> <script type="text/javascript" id="sage/main.js-js-extra"> /* <![CDATA[ */ var morePostsAjax = {"network_id":"1","site_url":"https:\/\/developer.nvidia.com\/blog\/wp-admin\/"}; /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/dist/scripts/main_61b387dd.js" id="sage/main.js-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/nv-recommender/js/dist/index.js" id="sage/recommender.js-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-includes/js/dist/vendor/lodash.min.js?ver=4.17.21" id="lodash-js"></script> <script type="text/javascript" id="lodash-js-after"> /* <![CDATA[ */ window.lodash = _.noConflict(); /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/ml-slider/assets/sliders/responsiveslides/responsiveslides.min.js" id="sage/flexslider.js-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-includes/js/comment-reply.min.js?ver=6.7.1" id="comment-reply-js" async="async" data-wp-strategy="async"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/themes/nvidia/resources/assets/libs/nv-developer-menu/nv-developer-menu.js?ver=1.0.0" id="nv-developer-menu.js-js"></script> <script type="text/javascript" id="stbClient-js-extra"> /* <![CDATA[ */ var stbUserOptions = {"restData":{"root":"https:\/\/developer-blogs.nvidia.com\/wp-json\/","nonce":"c8d2f6e0df"}}; /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/wp-special-textboxes/js/client.js?ver=6.2.1" id="stbClient-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/ml-slider/assets/sliders/flexslider/jquery.flexslider.min.js?ver=3.60.1" id="metaslider-flex-slider-js"></script> <script type="text/javascript" id="metaslider-flex-slider-js-after"> /* <![CDATA[ */ var metaslider_96052 = function($) { $('#metaslider_96052').flexslider({ slideshowSpeed:3000, animation:"fade", controlNav:false, directionNav:false, pauseOnHover:true, direction:"horizontal", reverse:false, keyboard:false, animationSpeed:600, prevText:"<", nextText:">", smoothHeight:false, fadeFirstSlide:false, slideshow:true }); $(document).trigger('metaslider/initialized', '#metaslider_96052'); }; var timer_metaslider_96052 = function() { var slider = !window.jQuery ? window.setTimeout(timer_metaslider_96052, 100) : !jQuery.isReady ? window.setTimeout(timer_metaslider_96052, 1) : metaslider_96052(window.jQuery); }; timer_metaslider_96052(); var metaslider_56284 = function($) { $('#metaslider_56284').flexslider({ slideshowSpeed:3000, animation:"fade", controlNav:false, directionNav:false, pauseOnHover:true, direction:"horizontal", reverse:false, keyboard:false, animationSpeed:600, prevText:"<", nextText:">", smoothHeight:false, fadeFirstSlide:false, slideshow:true }); $(document).trigger('metaslider/initialized', '#metaslider_56284'); }; var timer_metaslider_56284 = function() { var slider = !window.jQuery ? window.setTimeout(timer_metaslider_56284, 100) : !jQuery.isReady ? window.setTimeout(timer_metaslider_56284, 1) : metaslider_56284(window.jQuery); }; timer_metaslider_56284(); /* ]]> */ </script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/ml-slider/assets/metaslider/script.min.js?ver=3.60.1" id="metaslider-script-js"></script> <script type="text/javascript" src="https://developer-blogs.nvidia.com/wp-content/plugins/ml-slider/themes/highway/v1.0.0/script.js?ver=1.0.0" id="metaslider_highway_theme_script-js"></script> <script> function b2a(a){var b,c=0,l=0,f="",g=[];if(!a)return a;do{var e=a.charCodeAt(c++);var h=a.charCodeAt(c++);var k=a.charCodeAt(c++);var d=e<<16|h<<8|k;e=63&d>>18;h=63&d>>12;k=63&d>>6;d&=63;g[l++]="ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=".charAt(e)+"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=".charAt(h)+"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=".charAt(k)+"ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/=".charAt(d)}while(c< a.length);return f=g.join(""),b=a.length%3,(b?f.slice(0,b-3):f)+"===".slice(b||3)}function a2b(a){var b,c,l,f={},g=0,e=0,h="",k=String.fromCharCode,d=a.length;for(b=0;64>b;b++)f["ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/".charAt(b)]=b;for(c=0;d>c;c++)for(b=f[a.charAt(c)],g=(g<<6)+b,e+=6;8<=e;)((l=255&g>>>(e-=8))||d-2>c)&&(h+=k(l));return h}b64e=function(a){return btoa(encodeURIComponent(a).replace(/%([0-9A-F]{2})/g,function(b,a){return String.fromCharCode("0x"+a)}))}; b64d=function(a){return decodeURIComponent(atob(a).split("").map(function(a){return"%"+("00"+a.charCodeAt(0).toString(16)).slice(-2)}).join(""))}; /* <![CDATA[ */ ai_front = {"insertion_before":"BEFORE","insertion_after":"AFTER","insertion_prepend":"PREPEND CONTENT","insertion_append":"APPEND CONTENT","insertion_replace_content":"REPLACE CONTENT","insertion_replace_element":"REPLACE ELEMENT","visible":"VISIBLE","hidden":"HIDDEN","fallback":"FALLBACK","automatically_placed":"Automatically placed by AdSense Auto ads code","cancel":"Cancel","use":"Use","add":"Add","parent":"Parent","cancel_element_selection":"Cancel element selection","select_parent_element":"Select parent element","css_selector":"CSS selector","use_current_selector":"Use current selector","element":"ELEMENT","path":"PATH","selector":"SELECTOR"}; /* ]]> */ var ai_cookie_js=!0,ai_block_class_def="code-block"; /* JavaScript Cookie v2.2.0 https://github.com/js-cookie/js-cookie Copyright 2006, 2015 Klaus Hartl & Fagner Brack Released under the MIT license */ "undefined"!==typeof ai_cookie_js&&(function(a){if("function"===typeof define&&define.amd){define(a);var c=!0}"object"===typeof exports&&(module.exports=a(),c=!0);if(!c){var d=window.Cookies,b=window.Cookies=a();b.noConflict=function(){window.Cookies=d;return b}}}(function(){function a(){for(var d=0,b={};d<arguments.length;d++){var f=arguments[d],e;for(e in f)b[e]=f[e]}return b}function c(d){function b(){}function f(h,k,g){if("undefined"!==typeof document){g=a({path:"/",sameSite:"Lax"},b.defaults, g);"number"===typeof g.expires&&(g.expires=new Date(1*new Date+864E5*g.expires));g.expires=g.expires?g.expires.toUTCString():"";try{var l=JSON.stringify(k);/^[\{\[]/.test(l)&&(k=l)}catch(p){}k=d.write?d.write(k,h):encodeURIComponent(String(k)).replace(/%(23|24|26|2B|3A|3C|3E|3D|2F|3F|40|5B|5D|5E|60|7B|7D|7C)/g,decodeURIComponent);h=encodeURIComponent(String(h)).replace(/%(23|24|26|2B|5E|60|7C)/g,decodeURIComponent).replace(/[]/g,escape);l="";for(var n in g)g[n]&&(l+="; "+n,!0!==g[n]&&(l+="="+ g[n].split(";")[0]));return document.cookie=h+"="+k+l}}function e(h,k){if("undefined"!==typeof document){for(var g={},l=document.cookie?document.cookie.split("; "):[],n=0;n<l.length;n++){var p=l[n].split("="),m=p.slice(1).join("=");k||'"'!==m.charAt(0)||(m=m.slice(1,-1));try{var q=p[0].replace(/(%[0-9A-Z]{2})+/g,decodeURIComponent);m=(d.read||d)(m,q)||m.replace(/(%[0-9A-Z]{2})+/g,decodeURIComponent);if(k)try{m=JSON.parse(m)}catch(r){}g[q]=m;if(h===q)break}catch(r){}}return h?g[h]:g}}b.set=f;b.get= function(h){return e(h,!1)};b.getJSON=function(h){return e(h,!0)};b.remove=function(h,k){f(h,"",a(k,{expires:-1}))};b.defaults={};b.withConverter=c;return b}return c(function(){})}),AiCookies=Cookies.noConflict(),ai_check_block=function(a){if(null==a)return!0;var c=AiCookies.getJSON("aiBLOCKS");ai_debug_cookie_status="";null==c&&(c={});"undefined"!==typeof ai_delay_showing_pageviews&&(c.hasOwnProperty(a)||(c[a]={}),c[a].hasOwnProperty("d")||(c[a].d=ai_delay_showing_pageviews));if(c.hasOwnProperty(a)){for(var d in c[a]){if("x"== d){var b="",f=document.querySelectorAll('span[data-ai-block="'+a+'"]')[0];"aiHash"in f.dataset&&(b=f.dataset.aiHash);f="";c[a].hasOwnProperty("h")&&(f=c[a].h);var e=new Date;e=c[a][d]-Math.round(e.getTime()/1E3);if(0<e&&f==b)return ai_debug_cookie_status=a="closed for "+e+" s = "+Math.round(1E4*e/3600/24)/1E4+" days",!1;ai_set_cookie(a,"x","");c[a].hasOwnProperty("i")||c[a].hasOwnProperty("c")||ai_set_cookie(a,"h","")}else if("d"==d){if(0!=c[a][d])return ai_debug_cookie_status=a="delayed for "+c[a][d]+ " pageviews",!1}else if("i"==d){b="";f=document.querySelectorAll('span[data-ai-block="'+a+'"]')[0];"aiHash"in f.dataset&&(b=f.dataset.aiHash);f="";c[a].hasOwnProperty("h")&&(f=c[a].h);if(0==c[a][d]&&f==b)return ai_debug_cookie_status=a="max impressions reached",!1;if(0>c[a][d]&&f==b){e=new Date;e=-c[a][d]-Math.round(e.getTime()/1E3);if(0<e)return ai_debug_cookie_status=a="max imp. reached ("+Math.round(1E4*e/24/3600)/1E4+" days = "+e+" s)",!1;ai_set_cookie(a,"i","");c[a].hasOwnProperty("c")||c[a].hasOwnProperty("x")|| ai_set_cookie(a,"h","")}}if("ipt"==d&&0==c[a][d]&&(e=new Date,b=Math.round(e.getTime()/1E3),e=c[a].it-b,0<e))return ai_debug_cookie_status=a="max imp. per time reached ("+Math.round(1E4*e/24/3600)/1E4+" days = "+e+" s)",!1;if("c"==d){b="";f=document.querySelectorAll('span[data-ai-block="'+a+'"]')[0];"aiHash"in f.dataset&&(b=f.dataset.aiHash);f="";c[a].hasOwnProperty("h")&&(f=c[a].h);if(0==c[a][d]&&f==b)return ai_debug_cookie_status=a="max clicks reached",!1;if(0>c[a][d]&&f==b){e=new Date;e=-c[a][d]- Math.round(e.getTime()/1E3);if(0<e)return ai_debug_cookie_status=a="max clicks reached ("+Math.round(1E4*e/24/3600)/1E4+" days = "+e+" s)",!1;ai_set_cookie(a,"c","");c[a].hasOwnProperty("i")||c[a].hasOwnProperty("x")||ai_set_cookie(a,"h","")}}if("cpt"==d&&0==c[a][d]&&(e=new Date,b=Math.round(e.getTime()/1E3),e=c[a].ct-b,0<e))return ai_debug_cookie_status=a="max clicks per time reached ("+Math.round(1E4*e/24/3600)/1E4+" days = "+e+" s)",!1}if(c.hasOwnProperty("G")&&c.G.hasOwnProperty("cpt")&&0==c.G.cpt&& (e=new Date,b=Math.round(e.getTime()/1E3),e=c.G.ct-b,0<e))return ai_debug_cookie_status=a="max global clicks per time reached ("+Math.round(1E4*e/24/3600)/1E4+" days = "+e+" s)",!1}ai_debug_cookie_status="OK";return!0},ai_check_and_insert_block=function(a,c){if(null==a)return!0;var d=document.getElementsByClassName(c);if(d.length){d=d[0];var b=d.closest("."+ai_block_class_def),f=ai_check_block(a);!f&&0!=parseInt(d.getAttribute("limits-fallback"))&&d.hasAttribute("data-fallback-code")&&(d.setAttribute("data-code", d.getAttribute("data-fallback-code")),null!=b&&b.hasAttribute("data-ai")&&d.hasAttribute("fallback-tracking")&&d.hasAttribute("fallback_level")&&b.setAttribute("data-ai-"+d.getAttribute("fallback_level"),d.getAttribute("fallback-tracking")),f=!0);d.removeAttribute("data-selector");if(f)ai_insert_code(d),b&&(f=b.querySelectorAll(".ai-debug-block"),f.length&&(b.classList.remove("ai-list-block"),b.classList.remove("ai-list-block-ip"),b.classList.remove("ai-list-block-filter"),b.style.visibility="",b.classList.contains("ai-remove-position")&& (b.style.position="")));else{f=d.closest("div[data-ai]");if(null!=f&&"undefined"!=typeof f.getAttribute("data-ai")){var e=JSON.parse(b64d(f.getAttribute("data-ai")));"undefined"!==typeof e&&e.constructor===Array&&(e[1]="",f.setAttribute("data-ai",b64e(JSON.stringify(e))))}b&&(f=b.querySelectorAll(".ai-debug-block"),f.length&&(b.classList.remove("ai-list-block"),b.classList.remove("ai-list-block-ip"),b.classList.remove("ai-list-block-filter"),b.style.visibility="",b.classList.contains("ai-remove-position")&& (b.style.position="")))}d.classList.remove(c)}d=document.querySelectorAll("."+c+"-dbg");b=0;for(f=d.length;b<f;b++)e=d[b],e.querySelector(".ai-status").textContent=ai_debug_cookie_status,e.querySelector(".ai-cookie-data").textContent=ai_get_cookie_text(a),e.classList.remove(c+"-dbg")},ai_load_cookie=function(){var a=AiCookies.getJSON("aiBLOCKS");null==a&&(a={});return a},ai_set_cookie=function(a,c,d){var b=ai_load_cookie();if(""===d){if(b.hasOwnProperty(a)){delete b[a][c];a:{c=b[a];for(f in c)if(c.hasOwnProperty(f)){var f= !1;break a}f=!0}f&&delete b[a]}}else b.hasOwnProperty(a)||(b[a]={}),b[a][c]=d;0===Object.keys(b).length&&b.constructor===Object?AiCookies.remove("aiBLOCKS"):AiCookies.set("aiBLOCKS",b,{expires:365,path:"/"});return b},ai_get_cookie_text=function(a){var c=AiCookies.getJSON("aiBLOCKS");null==c&&(c={});var d="";c.hasOwnProperty("G")&&(d="G["+JSON.stringify(c.G).replace(/"/g,"").replace("{","").replace("}","")+"] ");var b="";c.hasOwnProperty(a)&&(b=JSON.stringify(c[a]).replace(/"/g,"").replace("{","").replace("}", ""));return d+b}); var ai_insertion_js=!0,ai_block_class_def="code-block"; if("undefined"!=typeof ai_insertion_js){ai_insert=function(a,h,l){if(-1!=h.indexOf(":eq("))if(window.jQuery&&window.jQuery.fn)var n=jQuery(h);else{console.error("AI INSERT USING jQuery QUERIES:",h,"- jQuery not found");return}else n=document.querySelectorAll(h);for(var u=0,y=n.length;u<y;u++){var d=n[u];selector_string=d.hasAttribute("id")?"#"+d.getAttribute("id"):d.hasAttribute("class")?"."+d.getAttribute("class").replace(RegExp(" ","g"),"."):"";var w=document.createElement("div");w.innerHTML=l; var m=w.getElementsByClassName("ai-selector-counter")[0];null!=m&&(m.innerText=u+1);m=w.getElementsByClassName("ai-debug-name ai-main")[0];if(null!=m){var r=a.toUpperCase();"undefined"!=typeof ai_front&&("before"==a?r=ai_front.insertion_before:"after"==a?r=ai_front.insertion_after:"prepend"==a?r=ai_front.insertion_prepend:"append"==a?r=ai_front.insertion_append:"replace-content"==a?r=ai_front.insertion_replace_content:"replace-element"==a&&(r=ai_front.insertion_replace_element));-1==selector_string.indexOf(".ai-viewports")&& (m.innerText=r+" "+h+" ("+d.tagName.toLowerCase()+selector_string+")")}m=document.createRange();try{var v=m.createContextualFragment(w.innerHTML)}catch(t){}"before"==a?d.parentNode.insertBefore(v,d):"after"==a?d.parentNode.insertBefore(v,d.nextSibling):"prepend"==a?d.insertBefore(v,d.firstChild):"append"==a?d.insertBefore(v,null):"replace-content"==a?(d.innerHTML="",d.insertBefore(v,null)):"replace-element"==a&&(d.parentNode.insertBefore(v,d),d.parentNode.removeChild(d));z()}};ai_insert_code=function(a){function h(m, r){return null==m?!1:m.classList?m.classList.contains(r):-1<(" "+m.className+" ").indexOf(" "+r+" ")}function l(m,r){null!=m&&(m.classList?m.classList.add(r):m.className+=" "+r)}function n(m,r){null!=m&&(m.classList?m.classList.remove(r):m.className=m.className.replace(new RegExp("(^|\\b)"+r.split(" ").join("|")+"(\\b|$)","gi")," "))}if("undefined"!=typeof a){var u=!1;if(h(a,"no-visibility-check")||a.offsetWidth||a.offsetHeight||a.getClientRects().length){u=a.getAttribute("data-code");var y=a.getAttribute("data-insertion-position"), d=a.getAttribute("data-selector");if(null!=u)if(null!=y&&null!=d){if(-1!=d.indexOf(":eq(")?window.jQuery&&window.jQuery.fn&&jQuery(d).length:document.querySelectorAll(d).length)ai_insert(y,d,b64d(u)),n(a,"ai-viewports")}else{y=document.createRange();try{var w=y.createContextualFragment(b64d(u))}catch(m){}a.parentNode.insertBefore(w,a.nextSibling);n(a,"ai-viewports")}u=!0}else w=a.previousElementSibling,h(w,"ai-debug-bar")&&h(w,"ai-debug-script")&&(n(w,"ai-debug-script"),l(w,"ai-debug-viewport-invisible")), n(a,"ai-viewports");return u}};ai_insert_list_code=function(a){var h=document.getElementsByClassName(a)[0];if("undefined"!=typeof h){var l=ai_insert_code(h),n=h.closest("div."+ai_block_class_def);if(n){l||n.removeAttribute("data-ai");var u=n.querySelectorAll(".ai-debug-block");n&&u.length&&(n.classList.remove("ai-list-block"),n.classList.remove("ai-list-block-ip"),n.classList.remove("ai-list-block-filter"),n.style.visibility="",n.classList.contains("ai-remove-position")&&(n.style.position=""))}h.classList.remove(a); l&&z()}};ai_insert_viewport_code=function(a){var h=document.getElementsByClassName(a)[0];if("undefined"!=typeof h){var l=ai_insert_code(h);h.classList.remove(a);l&&(a=h.closest("div."+ai_block_class_def),null!=a&&(l=h.getAttribute("style"),null!=l&&a.setAttribute("style",a.getAttribute("style")+" "+l)));setTimeout(function(){h.removeAttribute("style")},2);z()}};ai_insert_adsense_fallback_codes=function(a){a.style.display="none";var h=a.closest(".ai-fallback-adsense"),l=h.nextElementSibling;l.getAttribute("data-code")? ai_insert_code(l)&&z():l.style.display="block";h.classList.contains("ai-empty-code")&&null!=a.closest("."+ai_block_class_def)&&(a=a.closest("."+ai_block_class_def).getElementsByClassName("code-block-label"),0!=a.length&&(a[0].style.display="none"))};ai_insert_code_by_class=function(a){var h=document.getElementsByClassName(a)[0];"undefined"!=typeof h&&(ai_insert_code(h),h.classList.remove(a))};ai_insert_client_code=function(a,h){var l=document.getElementsByClassName(a)[0];if("undefined"!=typeof l){var n= l.getAttribute("data-code");null!=n&&ai_check_block()&&(l.setAttribute("data-code",n.substring(Math.floor(h/19))),ai_insert_code_by_class(a),l.remove())}};ai_process_elements_active=!1;function z(){ai_process_elements_active||setTimeout(function(){ai_process_elements_active=!1;"function"==typeof ai_process_rotations&&ai_process_rotations();"function"==typeof ai_process_lists&&ai_process_lists();"function"==typeof ai_process_ip_addresses&&ai_process_ip_addresses();"function"==typeof ai_process_filter_hooks&& ai_process_filter_hooks();"function"==typeof ai_adb_process_blocks&&ai_adb_process_blocks();"function"==typeof ai_process_impressions&&1==ai_tracking_finished&&ai_process_impressions();"function"==typeof ai_install_click_trackers&&1==ai_tracking_finished&&ai_install_click_trackers();"function"==typeof ai_install_close_buttons&&ai_install_close_buttons(document);"function"==typeof ai_process_wait_for_interaction&&ai_process_wait_for_interaction();"function"==typeof ai_process_delayed_blocks&&ai_process_delayed_blocks()}, 5);ai_process_elements_active=!0}const B=document.querySelector("body");(new MutationObserver(function(a,h){for(const l of a)"attributes"===l.type&&"data-ad-status"==l.attributeName&&"unfilled"==l.target.dataset.adStatus&&l.target.closest(".ai-fallback-adsense")&&ai_insert_adsense_fallback_codes(l.target)})).observe(B,{attributes:!0,childList:!1,subtree:!0});var Arrive=function(a,h,l){function n(t,c,e){d.addMethod(c,e,t.unbindEvent);d.addMethod(c,e,t.unbindEventWithSelectorOrCallback);d.addMethod(c, e,t.unbindEventWithSelectorAndCallback)}function u(t){t.arrive=r.bindEvent;n(r,t,"unbindArrive");t.leave=v.bindEvent;n(v,t,"unbindLeave")}if(a.MutationObserver&&"undefined"!==typeof HTMLElement){var y=0,d=function(){var t=HTMLElement.prototype.matches||HTMLElement.prototype.webkitMatchesSelector||HTMLElement.prototype.mozMatchesSelector||HTMLElement.prototype.msMatchesSelector;return{matchesSelector:function(c,e){return c instanceof HTMLElement&&t.call(c,e)},addMethod:function(c,e,f){var b=c[e];c[e]= function(){if(f.length==arguments.length)return f.apply(this,arguments);if("function"==typeof b)return b.apply(this,arguments)}},callCallbacks:function(c,e){e&&e.options.onceOnly&&1==e.firedElems.length&&(c=[c[0]]);for(var f=0,b;b=c[f];f++)b&&b.callback&&b.callback.call(b.elem,b.elem);e&&e.options.onceOnly&&1==e.firedElems.length&&e.me.unbindEventWithSelectorAndCallback.call(e.target,e.selector,e.callback)},checkChildNodesRecursively:function(c,e,f,b){for(var g=0,k;k=c[g];g++)f(k,e,b)&&b.push({callback:e.callback, elem:k}),0<k.childNodes.length&&d.checkChildNodesRecursively(k.childNodes,e,f,b)},mergeArrays:function(c,e){var f={},b;for(b in c)c.hasOwnProperty(b)&&(f[b]=c[b]);for(b in e)e.hasOwnProperty(b)&&(f[b]=e[b]);return f},toElementsArray:function(c){"undefined"===typeof c||"number"===typeof c.length&&c!==a||(c=[c]);return c}}}(),w=function(){var t=function(){this._eventsBucket=[];this._beforeRemoving=this._beforeAdding=null};t.prototype.addEvent=function(c,e,f,b){c={target:c,selector:e,options:f,callback:b, firedElems:[]};this._beforeAdding&&this._beforeAdding(c);this._eventsBucket.push(c);return c};t.prototype.removeEvent=function(c){for(var e=this._eventsBucket.length-1,f;f=this._eventsBucket[e];e--)c(f)&&(this._beforeRemoving&&this._beforeRemoving(f),(f=this._eventsBucket.splice(e,1))&&f.length&&(f[0].callback=null))};t.prototype.beforeAdding=function(c){this._beforeAdding=c};t.prototype.beforeRemoving=function(c){this._beforeRemoving=c};return t}(),m=function(t,c){var e=new w,f=this,b={fireOnAttributesModification:!1}; e.beforeAdding(function(g){var k=g.target;if(k===a.document||k===a)k=document.getElementsByTagName("html")[0];var p=new MutationObserver(function(x){c.call(this,x,g)});var q=t(g.options);p.observe(k,q);g.observer=p;g.me=f});e.beforeRemoving(function(g){g.observer.disconnect()});this.bindEvent=function(g,k,p){k=d.mergeArrays(b,k);for(var q=d.toElementsArray(this),x=0;x<q.length;x++)e.addEvent(q[x],g,k,p)};this.unbindEvent=function(){var g=d.toElementsArray(this);e.removeEvent(function(k){for(var p= 0;p<g.length;p++)if(this===l||k.target===g[p])return!0;return!1})};this.unbindEventWithSelectorOrCallback=function(g){var k=d.toElementsArray(this);e.removeEvent("function"===typeof g?function(p){for(var q=0;q<k.length;q++)if((this===l||p.target===k[q])&&p.callback===g)return!0;return!1}:function(p){for(var q=0;q<k.length;q++)if((this===l||p.target===k[q])&&p.selector===g)return!0;return!1})};this.unbindEventWithSelectorAndCallback=function(g,k){var p=d.toElementsArray(this);e.removeEvent(function(q){for(var x= 0;x<p.length;x++)if((this===l||q.target===p[x])&&q.selector===g&&q.callback===k)return!0;return!1})};return this},r=new function(){function t(f,b,g){return d.matchesSelector(f,b.selector)&&(f._id===l&&(f._id=y++),-1==b.firedElems.indexOf(f._id))?(b.firedElems.push(f._id),!0):!1}var c={fireOnAttributesModification:!1,onceOnly:!1,existing:!1};r=new m(function(f){var b={attributes:!1,childList:!0,subtree:!0};f.fireOnAttributesModification&&(b.attributes=!0);return b},function(f,b){f.forEach(function(g){var k= g.addedNodes,p=g.target,q=[];null!==k&&0<k.length?d.checkChildNodesRecursively(k,b,t,q):"attributes"===g.type&&t(p,b,q)&&q.push({callback:b.callback,elem:p});d.callCallbacks(q,b)})});var e=r.bindEvent;r.bindEvent=function(f,b,g){"undefined"===typeof g?(g=b,b=c):b=d.mergeArrays(c,b);var k=d.toElementsArray(this);if(b.existing){for(var p=[],q=0;q<k.length;q++)for(var x=k[q].querySelectorAll(f),A=0;A<x.length;A++)p.push({callback:g,elem:x[A]});if(b.onceOnly&&p.length)return g.call(p[0].elem,p[0].elem); setTimeout(d.callCallbacks,1,p)}e.call(this,f,b,g)};return r},v=new function(){function t(f,b){return d.matchesSelector(f,b.selector)}var c={};v=new m(function(){return{childList:!0,subtree:!0}},function(f,b){f.forEach(function(g){g=g.removedNodes;var k=[];null!==g&&0<g.length&&d.checkChildNodesRecursively(g,b,t,k);d.callCallbacks(k,b)})});var e=v.bindEvent;v.bindEvent=function(f,b,g){"undefined"===typeof g?(g=b,b=c):b=d.mergeArrays(c,b);e.call(this,f,b,g)};return v};h&&u(h.fn);u(HTMLElement.prototype); u(NodeList.prototype);u(HTMLCollection.prototype);u(HTMLDocument.prototype);u(Window.prototype);h={};n(r,h,"unbindAllArrive");n(v,h,"unbindAllLeave");return h}}(window,"undefined"===typeof jQuery?null:jQuery,void 0)}; var ai_rotation_triggers=[],ai_block_class_def="code-block"; if("undefined"!=typeof ai_rotation_triggers){ai_process_rotation=function(b){var d="number"==typeof b.length;window.jQuery&&window.jQuery.fn&&b instanceof jQuery&&(b=d?Array.prototype.slice.call(b):b[0]);if(d){var e=!1;b.forEach((c,h)=>{if(c.classList.contains("ai-unprocessed")||c.classList.contains("ai-timer"))e=!0});if(!e)return;b.forEach((c,h)=>{c.classList.remove("ai-unprocessed");c.classList.remove("ai-timer")})}else{if(!b.classList.contains("ai-unprocessed")&&!b.classList.contains("ai-timer"))return; b.classList.remove("ai-unprocessed");b.classList.remove("ai-timer")}var a=!1;if(d?b[0].hasAttribute("data-info"):b.hasAttribute("data-info")){var f="div.ai-rotate.ai-"+(d?JSON.parse(atob(b[0].dataset.info)):JSON.parse(atob(b.dataset.info)))[0];ai_rotation_triggers.includes(f)&&(ai_rotation_triggers.splice(ai_rotation_triggers.indexOf(f),1),a=!0)}if(d)for(d=0;d<b.length;d++)0==d?ai_process_single_rotation(b[d],!0):ai_process_single_rotation(b[d],!1);else ai_process_single_rotation(b,!a)};ai_process_single_rotation= function(b,d){var e=[];Array.from(b.children).forEach((g,p)=>{g.matches(".ai-rotate-option")&&e.push(g)});if(0!=e.length){e.forEach((g,p)=>{g.style.display="none"});if(b.hasAttribute("data-next")){k=parseInt(b.getAttribute("data-next"));var a=e[k];if(a.hasAttribute("data-code")){var f=document.createRange(),c=!0;try{var h=f.createContextualFragment(b64d(a.dataset.code))}catch(g){c=!1}c&&(a=h)}0!=a.querySelectorAll("span[data-ai-groups]").length&&0!=document.querySelectorAll(".ai-rotation-groups").length&& setTimeout(function(){B()},5)}else if(e[0].hasAttribute("data-group")){var k=-1,u=[];document.querySelectorAll("span[data-ai-groups]").forEach((g,p)=>{(g.offsetWidth||g.offsetHeight||g.getClientRects().length)&&u.push(g)});1<=u.length&&(timed_groups=[],groups=[],u.forEach(function(g,p){active_groups=JSON.parse(b64d(g.dataset.aiGroups));var r=!1;g=g.closest(".ai-rotate");null!=g&&g.classList.contains("ai-timed-rotation")&&(r=!0);active_groups.forEach(function(t,v){groups.push(t);r&&timed_groups.push(t)})}), groups.forEach(function(g,p){-1==k&&e.forEach((r,t)=>{var v=b64d(r.dataset.group);option_group_items=v.split(",");option_group_items.forEach(function(C,E){-1==k&&C.trim()==g&&(k=t,timed_groups.includes(v)&&b.classList.add("ai-timed-rotation"))})})}))}else if(b.hasAttribute("data-shares"))for(f=JSON.parse(atob(b.dataset.shares)),a=Math.round(100*Math.random()),c=0;c<f.length&&(k=c,0>f[c]||!(a<=f[c]));c++);else f=b.classList.contains("ai-unique"),a=new Date,f?("number"!=typeof ai_rotation_seed&&(ai_rotation_seed= (Math.floor(1E3*Math.random())+a.getMilliseconds())%e.length),f=ai_rotation_seed,f>e.length&&(f%=e.length),a=parseInt(b.dataset.counter),a<=e.length?(k=parseInt(f+a-1),k>=e.length&&(k-=e.length)):k=e.length):(k=Math.floor(Math.random()*e.length),a.getMilliseconds()%2&&(k=e.length-k-1));if(b.classList.contains("ai-rotation-scheduling"))for(k=-1,f=0;f<e.length;f++)if(a=e[f],a.hasAttribute("data-scheduling")){c=b64d(a.dataset.scheduling);a=!0;0==c.indexOf("^")&&(a=!1,c=c.substring(1));var q=c.split("="), m=-1!=c.indexOf("%")?q[0].split("%"):[q[0]];c=m[0].trim().toLowerCase();m="undefined"!=typeof m[1]?m[1].trim():0;q=q[1].replace(" ","");var n=(new Date).getTime();n=new Date(n);var l=0;switch(c){case "s":l=n.getSeconds();break;case "i":l=n.getMinutes();break;case "h":l=n.getHours();break;case "d":l=n.getDate();break;case "m":l=n.getMonth();break;case "y":l=n.getFullYear();break;case "w":l=n.getDay(),l=0==l?6:l-1}c=0!=m?l%m:l;m=q.split(",");q=!a;for(n=0;n<m.length;n++)if(l=m[n],-1!=l.indexOf("-")){if(l= l.split("-"),c>=l[0]&&c<=l[1]){q=a;break}}else if(c==l){q=a;break}if(q){k=f;break}}if(!(0>k||k>=e.length)){a=e[k];var z="",w=b.classList.contains("ai-timed-rotation");e.forEach((g,p)=>{g.hasAttribute("data-time")&&(w=!0)});if(a.hasAttribute("data-time")){f=atob(a.dataset.time);if(0==f&&1<e.length){c=k;do{c++;c>=e.length&&(c=0);m=e[c];if(!m.hasAttribute("data-time")){k=c;a=e[k];f=0;break}m=atob(m.dataset.time)}while(0==m&&c!=k);0!=f&&(k=c,a=e[k],f=atob(a.dataset.time))}if(0<f&&(c=k+1,c>=e.length&& (c=0),b.hasAttribute("data-info"))){m=JSON.parse(atob(b.dataset.info))[0];b.setAttribute("data-next",c);var x="div.ai-rotate.ai-"+m;ai_rotation_triggers.includes(x)&&(d=!1);d&&(ai_rotation_triggers.push(x),setTimeout(function(){var g=document.querySelectorAll(x);g.forEach((p,r)=>{p.classList.add("ai-timer")});ai_process_rotation(g)},1E3*f));z=" ("+f+" s)"}}else a.hasAttribute("data-group")||e.forEach((g,p)=>{p!=k&&g.remove()});a.style.display="";a.style.visibility="";a.style.position="";a.style.width= "";a.style.height="";a.style.top="";a.style.left="";a.classList.remove("ai-rotate-hidden");a.classList.remove("ai-rotate-hidden-2");b.style.position="";if(a.hasAttribute("data-code")){e.forEach((g,p)=>{g.innerText=""});d=b64d(a.dataset.code);f=document.createRange();c=!0;try{h=f.createContextualFragment(d)}catch(g){c=!1}a.append(h);D()}f=parseInt(a.dataset.index);var y=b64d(a.dataset.name);d=b.closest(".ai-debug-block");if(null!=d){h=d.querySelectorAll("kbd.ai-option-name");d=d.querySelectorAll(".ai-debug-block"); if(0!=d.length){var A=[];d.forEach((g,p)=>{g.querySelectorAll("kbd.ai-option-name").forEach((r,t)=>{A.push(r)})});h=Array.from(h);h=h.slice(0,h.length-A.length)}0!=h.length&&(separator=h[0].hasAttribute("data-separator")?h[0].dataset.separator:"",h.forEach((g,p)=>{g.innerText=separator+y+z}))}d=!1;a=b.closest(".ai-adb-show");null!=a&&a.hasAttribute("data-ai-tracking")&&(h=JSON.parse(b64d(a.getAttribute("data-ai-tracking"))),"undefined"!==typeof h&&h.constructor===Array&&(h[1]=f,h[3]=y,a.setAttribute("data-ai-tracking", b64e(JSON.stringify(h))),a.classList.add("ai-track"),w&&ai_tracking_finished&&a.classList.add("ai-no-pageview"),d=!0));d||(d=b.closest("div[data-ai]"),null!=d&&d.hasAttribute("data-ai")&&(h=JSON.parse(b64d(d.getAttribute("data-ai"))),"undefined"!==typeof h&&h.constructor===Array&&(h[1]=f,h[3]=y,d.setAttribute("data-ai",b64e(JSON.stringify(h))),d.classList.add("ai-track"),w&&ai_tracking_finished&&d.classList.add("ai-no-pageview"))))}}};ai_process_rotations=function(){document.querySelectorAll("div.ai-rotate").forEach((b, d)=>{ai_process_rotation(b)})};function B(){document.querySelectorAll("div.ai-rotate.ai-rotation-groups").forEach((b,d)=>{b.classList.add("ai-timer");ai_process_rotation(b)})}ai_process_rotations_in_element=function(b){null!=b&&b.querySelectorAll("div.ai-rotate").forEach((d,e)=>{ai_process_rotation(d)})};(function(b){"complete"===document.readyState||"loading"!==document.readyState&&!document.documentElement.doScroll?b():document.addEventListener("DOMContentLoaded",b)})(function(){setTimeout(function(){ai_process_rotations()}, 10)});ai_process_elements_active=!1;function D(){ai_process_elements_active||setTimeout(function(){ai_process_elements_active=!1;"function"==typeof ai_process_rotations&&ai_process_rotations();"function"==typeof ai_process_lists&&ai_process_lists();"function"==typeof ai_process_ip_addresses&&ai_process_ip_addresses();"function"==typeof ai_process_filter_hooks&&ai_process_filter_hooks();"function"==typeof ai_adb_process_blocks&&ai_adb_process_blocks();"function"==typeof ai_process_impressions&&1== ai_tracking_finished&&ai_process_impressions();"function"==typeof ai_install_click_trackers&&1==ai_tracking_finished&&ai_install_click_trackers();"function"==typeof ai_install_close_buttons&&ai_install_close_buttons(document)},5);ai_process_elements_active=!0}}; ;!function(a,b){a(function(){"use strict";function a(a,b){return null!=a&&null!=b&&a.toLowerCase()===b.toLowerCase()}function c(a,b){var c,d,e=a.length;if(!e||!b)return!1;for(c=b.toLowerCase(),d=0;d<e;++d)if(c===a[d].toLowerCase())return!0;return!1}function d(a){for(var b in a)i.call(a,b)&&(a[b]=new RegExp(a[b],"i"))}function e(a){return(a||"").substr(0,500)}function f(a,b){this.ua=e(a),this._cache={},this.maxPhoneWidth=b||600}var g={};g.mobileDetectRules={phones:{iPhone:"\\biPhone\\b|\\biPod\\b",BlackBerry:"BlackBerry|\\bBB10\\b|rim[0-9]+|\\b(BBA100|BBB100|BBD100|BBE100|BBF100|STH100)\\b-[0-9]+",Pixel:"; \\bPixel\\b",HTC:"HTC|HTC.*(Sensation|Evo|Vision|Explorer|6800|8100|8900|A7272|S510e|C110e|Legend|Desire|T8282)|APX515CKT|Qtek9090|APA9292KT|HD_mini|Sensation.*Z710e|PG86100|Z715e|Desire.*(A8181|HD)|ADR6200|ADR6400L|ADR6425|001HT|Inspire 4G|Android.*\\bEVO\\b|T-Mobile G1|Z520m|Android [0-9.]+; Pixel",Nexus:"Nexus One|Nexus S|Galaxy.*Nexus|Android.*Nexus.*Mobile|Nexus 4|Nexus 5|Nexus 5X|Nexus 6",Dell:"Dell[;]? (Streak|Aero|Venue|Venue Pro|Flash|Smoke|Mini 3iX)|XCD28|XCD35|\\b001DL\\b|\\b101DL\\b|\\bGS01\\b",Motorola:"Motorola|DROIDX|DROID BIONIC|\\bDroid\\b.*Build|Android.*Xoom|HRI39|MOT-|A1260|A1680|A555|A853|A855|A953|A955|A956|Motorola.*ELECTRIFY|Motorola.*i1|i867|i940|MB200|MB300|MB501|MB502|MB508|MB511|MB520|MB525|MB526|MB611|MB612|MB632|MB810|MB855|MB860|MB861|MB865|MB870|ME501|ME502|ME511|ME525|ME600|ME632|ME722|ME811|ME860|ME863|ME865|MT620|MT710|MT716|MT720|MT810|MT870|MT917|Motorola.*TITANIUM|WX435|WX445|XT300|XT301|XT311|XT316|XT317|XT319|XT320|XT390|XT502|XT530|XT531|XT532|XT535|XT603|XT610|XT611|XT615|XT681|XT701|XT702|XT711|XT720|XT800|XT806|XT860|XT862|XT875|XT882|XT883|XT894|XT901|XT907|XT909|XT910|XT912|XT928|XT926|XT915|XT919|XT925|XT1021|\\bMoto E\\b|XT1068|XT1092|XT1052",Samsung:"\\bSamsung\\b|SM-G950F|SM-G955F|SM-G9250|GT-19300|SGH-I337|BGT-S5230|GT-B2100|GT-B2700|GT-B2710|GT-B3210|GT-B3310|GT-B3410|GT-B3730|GT-B3740|GT-B5510|GT-B5512|GT-B5722|GT-B6520|GT-B7300|GT-B7320|GT-B7330|GT-B7350|GT-B7510|GT-B7722|GT-B7800|GT-C3010|GT-C3011|GT-C3060|GT-C3200|GT-C3212|GT-C3212I|GT-C3262|GT-C3222|GT-C3300|GT-C3300K|GT-C3303|GT-C3303K|GT-C3310|GT-C3322|GT-C3330|GT-C3350|GT-C3500|GT-C3510|GT-C3530|GT-C3630|GT-C3780|GT-C5010|GT-C5212|GT-C6620|GT-C6625|GT-C6712|GT-E1050|GT-E1070|GT-E1075|GT-E1080|GT-E1081|GT-E1085|GT-E1087|GT-E1100|GT-E1107|GT-E1110|GT-E1120|GT-E1125|GT-E1130|GT-E1160|GT-E1170|GT-E1175|GT-E1180|GT-E1182|GT-E1200|GT-E1210|GT-E1225|GT-E1230|GT-E1390|GT-E2100|GT-E2120|GT-E2121|GT-E2152|GT-E2220|GT-E2222|GT-E2230|GT-E2232|GT-E2250|GT-E2370|GT-E2550|GT-E2652|GT-E3210|GT-E3213|GT-I5500|GT-I5503|GT-I5700|GT-I5800|GT-I5801|GT-I6410|GT-I6420|GT-I7110|GT-I7410|GT-I7500|GT-I8000|GT-I8150|GT-I8160|GT-I8190|GT-I8320|GT-I8330|GT-I8350|GT-I8530|GT-I8700|GT-I8703|GT-I8910|GT-I9000|GT-I9001|GT-I9003|GT-I9010|GT-I9020|GT-I9023|GT-I9070|GT-I9082|GT-I9100|GT-I9103|GT-I9220|GT-I9250|GT-I9300|GT-I9305|GT-I9500|GT-I9505|GT-M3510|GT-M5650|GT-M7500|GT-M7600|GT-M7603|GT-M8800|GT-M8910|GT-N7000|GT-S3110|GT-S3310|GT-S3350|GT-S3353|GT-S3370|GT-S3650|GT-S3653|GT-S3770|GT-S3850|GT-S5210|GT-S5220|GT-S5229|GT-S5230|GT-S5233|GT-S5250|GT-S5253|GT-S5260|GT-S5263|GT-S5270|GT-S5300|GT-S5330|GT-S5350|GT-S5360|GT-S5363|GT-S5369|GT-S5380|GT-S5380D|GT-S5560|GT-S5570|GT-S5600|GT-S5603|GT-S5610|GT-S5620|GT-S5660|GT-S5670|GT-S5690|GT-S5750|GT-S5780|GT-S5830|GT-S5839|GT-S6102|GT-S6500|GT-S7070|GT-S7200|GT-S7220|GT-S7230|GT-S7233|GT-S7250|GT-S7500|GT-S7530|GT-S7550|GT-S7562|GT-S7710|GT-S8000|GT-S8003|GT-S8500|GT-S8530|GT-S8600|SCH-A310|SCH-A530|SCH-A570|SCH-A610|SCH-A630|SCH-A650|SCH-A790|SCH-A795|SCH-A850|SCH-A870|SCH-A890|SCH-A930|SCH-A950|SCH-A970|SCH-A990|SCH-I100|SCH-I110|SCH-I400|SCH-I405|SCH-I500|SCH-I510|SCH-I515|SCH-I600|SCH-I730|SCH-I760|SCH-I770|SCH-I830|SCH-I910|SCH-I920|SCH-I959|SCH-LC11|SCH-N150|SCH-N300|SCH-R100|SCH-R300|SCH-R351|SCH-R400|SCH-R410|SCH-T300|SCH-U310|SCH-U320|SCH-U350|SCH-U360|SCH-U365|SCH-U370|SCH-U380|SCH-U410|SCH-U430|SCH-U450|SCH-U460|SCH-U470|SCH-U490|SCH-U540|SCH-U550|SCH-U620|SCH-U640|SCH-U650|SCH-U660|SCH-U700|SCH-U740|SCH-U750|SCH-U810|SCH-U820|SCH-U900|SCH-U940|SCH-U960|SCS-26UC|SGH-A107|SGH-A117|SGH-A127|SGH-A137|SGH-A157|SGH-A167|SGH-A177|SGH-A187|SGH-A197|SGH-A227|SGH-A237|SGH-A257|SGH-A437|SGH-A517|SGH-A597|SGH-A637|SGH-A657|SGH-A667|SGH-A687|SGH-A697|SGH-A707|SGH-A717|SGH-A727|SGH-A737|SGH-A747|SGH-A767|SGH-A777|SGH-A797|SGH-A817|SGH-A827|SGH-A837|SGH-A847|SGH-A867|SGH-A877|SGH-A887|SGH-A897|SGH-A927|SGH-B100|SGH-B130|SGH-B200|SGH-B220|SGH-C100|SGH-C110|SGH-C120|SGH-C130|SGH-C140|SGH-C160|SGH-C170|SGH-C180|SGH-C200|SGH-C207|SGH-C210|SGH-C225|SGH-C230|SGH-C417|SGH-C450|SGH-D307|SGH-D347|SGH-D357|SGH-D407|SGH-D415|SGH-D780|SGH-D807|SGH-D980|SGH-E105|SGH-E200|SGH-E315|SGH-E316|SGH-E317|SGH-E335|SGH-E590|SGH-E635|SGH-E715|SGH-E890|SGH-F300|SGH-F480|SGH-I200|SGH-I300|SGH-I320|SGH-I550|SGH-I577|SGH-I600|SGH-I607|SGH-I617|SGH-I627|SGH-I637|SGH-I677|SGH-I700|SGH-I717|SGH-I727|SGH-i747M|SGH-I777|SGH-I780|SGH-I827|SGH-I847|SGH-I857|SGH-I896|SGH-I897|SGH-I900|SGH-I907|SGH-I917|SGH-I927|SGH-I937|SGH-I997|SGH-J150|SGH-J200|SGH-L170|SGH-L700|SGH-M110|SGH-M150|SGH-M200|SGH-N105|SGH-N500|SGH-N600|SGH-N620|SGH-N625|SGH-N700|SGH-N710|SGH-P107|SGH-P207|SGH-P300|SGH-P310|SGH-P520|SGH-P735|SGH-P777|SGH-Q105|SGH-R210|SGH-R220|SGH-R225|SGH-S105|SGH-S307|SGH-T109|SGH-T119|SGH-T139|SGH-T209|SGH-T219|SGH-T229|SGH-T239|SGH-T249|SGH-T259|SGH-T309|SGH-T319|SGH-T329|SGH-T339|SGH-T349|SGH-T359|SGH-T369|SGH-T379|SGH-T409|SGH-T429|SGH-T439|SGH-T459|SGH-T469|SGH-T479|SGH-T499|SGH-T509|SGH-T519|SGH-T539|SGH-T559|SGH-T589|SGH-T609|SGH-T619|SGH-T629|SGH-T639|SGH-T659|SGH-T669|SGH-T679|SGH-T709|SGH-T719|SGH-T729|SGH-T739|SGH-T746|SGH-T749|SGH-T759|SGH-T769|SGH-T809|SGH-T819|SGH-T839|SGH-T919|SGH-T929|SGH-T939|SGH-T959|SGH-T989|SGH-U100|SGH-U200|SGH-U800|SGH-V205|SGH-V206|SGH-X100|SGH-X105|SGH-X120|SGH-X140|SGH-X426|SGH-X427|SGH-X475|SGH-X495|SGH-X497|SGH-X507|SGH-X600|SGH-X610|SGH-X620|SGH-X630|SGH-X700|SGH-X820|SGH-X890|SGH-Z130|SGH-Z150|SGH-Z170|SGH-ZX10|SGH-ZX20|SHW-M110|SPH-A120|SPH-A400|SPH-A420|SPH-A460|SPH-A500|SPH-A560|SPH-A600|SPH-A620|SPH-A660|SPH-A700|SPH-A740|SPH-A760|SPH-A790|SPH-A800|SPH-A820|SPH-A840|SPH-A880|SPH-A900|SPH-A940|SPH-A960|SPH-D600|SPH-D700|SPH-D710|SPH-D720|SPH-I300|SPH-I325|SPH-I330|SPH-I350|SPH-I500|SPH-I600|SPH-I700|SPH-L700|SPH-M100|SPH-M220|SPH-M240|SPH-M300|SPH-M305|SPH-M320|SPH-M330|SPH-M350|SPH-M360|SPH-M370|SPH-M380|SPH-M510|SPH-M540|SPH-M550|SPH-M560|SPH-M570|SPH-M580|SPH-M610|SPH-M620|SPH-M630|SPH-M800|SPH-M810|SPH-M850|SPH-M900|SPH-M910|SPH-M920|SPH-M930|SPH-N100|SPH-N200|SPH-N240|SPH-N300|SPH-N400|SPH-Z400|SWC-E100|SCH-i909|GT-N7100|GT-N7105|SCH-I535|SM-N900A|SGH-I317|SGH-T999L|GT-S5360B|GT-I8262|GT-S6802|GT-S6312|GT-S6310|GT-S5312|GT-S5310|GT-I9105|GT-I8510|GT-S6790N|SM-G7105|SM-N9005|GT-S5301|GT-I9295|GT-I9195|SM-C101|GT-S7392|GT-S7560|GT-B7610|GT-I5510|GT-S7582|GT-S7530E|GT-I8750|SM-G9006V|SM-G9008V|SM-G9009D|SM-G900A|SM-G900D|SM-G900F|SM-G900H|SM-G900I|SM-G900J|SM-G900K|SM-G900L|SM-G900M|SM-G900P|SM-G900R4|SM-G900S|SM-G900T|SM-G900V|SM-G900W8|SHV-E160K|SCH-P709|SCH-P729|SM-T2558|GT-I9205|SM-G9350|SM-J120F|SM-G920F|SM-G920V|SM-G930F|SM-N910C|SM-A310F|GT-I9190|SM-J500FN|SM-G903F|SM-J330F|SM-G610F|SM-G981B|SM-G892A|SM-A530F",LG:"\\bLG\\b;|LG[- ]?(C800|C900|E400|E610|E900|E-900|F160|F180K|F180L|F180S|730|855|L160|LS740|LS840|LS970|LU6200|MS690|MS695|MS770|MS840|MS870|MS910|P500|P700|P705|VM696|AS680|AS695|AX840|C729|E970|GS505|272|C395|E739BK|E960|L55C|L75C|LS696|LS860|P769BK|P350|P500|P509|P870|UN272|US730|VS840|VS950|LN272|LN510|LS670|LS855|LW690|MN270|MN510|P509|P769|P930|UN200|UN270|UN510|UN610|US670|US740|US760|UX265|UX840|VN271|VN530|VS660|VS700|VS740|VS750|VS910|VS920|VS930|VX9200|VX11000|AX840A|LW770|P506|P925|P999|E612|D955|D802|MS323|M257)|LM-G710",Sony:"SonyST|SonyLT|SonyEricsson|SonyEricssonLT15iv|LT18i|E10i|LT28h|LT26w|SonyEricssonMT27i|C5303|C6902|C6903|C6906|C6943|D2533|SOV34|601SO|F8332",Asus:"Asus.*Galaxy|PadFone.*Mobile",Xiaomi:"^(?!.*\\bx11\\b).*xiaomi.*$|POCOPHONE F1|MI 8|Redmi Note 9S|Redmi Note 5A Prime|N2G47H|M2001J2G|M2001J2I|M1805E10A|M2004J11G|M1902F1G|M2002J9G|M2004J19G|M2003J6A1G",NokiaLumia:"Lumia [0-9]{3,4}",Micromax:"Micromax.*\\b(A210|A92|A88|A72|A111|A110Q|A115|A116|A110|A90S|A26|A51|A35|A54|A25|A27|A89|A68|A65|A57|A90)\\b",Palm:"PalmSource|Palm",Vertu:"Vertu|Vertu.*Ltd|Vertu.*Ascent|Vertu.*Ayxta|Vertu.*Constellation(F|Quest)?|Vertu.*Monika|Vertu.*Signature",Pantech:"PANTECH|IM-A850S|IM-A840S|IM-A830L|IM-A830K|IM-A830S|IM-A820L|IM-A810K|IM-A810S|IM-A800S|IM-T100K|IM-A725L|IM-A780L|IM-A775C|IM-A770K|IM-A760S|IM-A750K|IM-A740S|IM-A730S|IM-A720L|IM-A710K|IM-A690L|IM-A690S|IM-A650S|IM-A630K|IM-A600S|VEGA PTL21|PT003|P8010|ADR910L|P6030|P6020|P9070|P4100|P9060|P5000|CDM8992|TXT8045|ADR8995|IS11PT|P2030|P6010|P8000|PT002|IS06|CDM8999|P9050|PT001|TXT8040|P2020|P9020|P2000|P7040|P7000|C790",Fly:"IQ230|IQ444|IQ450|IQ440|IQ442|IQ441|IQ245|IQ256|IQ236|IQ255|IQ235|IQ245|IQ275|IQ240|IQ285|IQ280|IQ270|IQ260|IQ250",Wiko:"KITE 4G|HIGHWAY|GETAWAY|STAIRWAY|DARKSIDE|DARKFULL|DARKNIGHT|DARKMOON|SLIDE|WAX 4G|RAINBOW|BLOOM|SUNSET|GOA(?!nna)|LENNY|BARRY|IGGY|OZZY|CINK FIVE|CINK PEAX|CINK PEAX 2|CINK SLIM|CINK SLIM 2|CINK +|CINK KING|CINK PEAX|CINK SLIM|SUBLIM",iMobile:"i-mobile (IQ|i-STYLE|idea|ZAA|Hitz)",SimValley:"\\b(SP-80|XT-930|SX-340|XT-930|SX-310|SP-360|SP60|SPT-800|SP-120|SPT-800|SP-140|SPX-5|SPX-8|SP-100|SPX-8|SPX-12)\\b",Wolfgang:"AT-B24D|AT-AS50HD|AT-AS40W|AT-AS55HD|AT-AS45q2|AT-B26D|AT-AS50Q",Alcatel:"Alcatel",Nintendo:"Nintendo (3DS|Switch)",Amoi:"Amoi",INQ:"INQ",OnePlus:"ONEPLUS",GenericPhone:"Tapatalk|PDA;|SAGEM|\\bmmp\\b|pocket|\\bpsp\\b|symbian|Smartphone|smartfon|treo|up.browser|up.link|vodafone|\\bwap\\b|nokia|Series40|Series60|S60|SonyEricsson|N900|MAUI.*WAP.*Browser"},tablets:{iPad:"iPad|iPad.*Mobile",NexusTablet:"Android.*Nexus[\\s]+(7|9|10)",GoogleTablet:"Android.*Pixel C",SamsungTablet:"SAMSUNG.*Tablet|Galaxy.*Tab|SC-01C|GT-P1000|GT-P1003|GT-P1010|GT-P3105|GT-P6210|GT-P6800|GT-P6810|GT-P7100|GT-P7300|GT-P7310|GT-P7500|GT-P7510|SCH-I800|SCH-I815|SCH-I905|SGH-I957|SGH-I987|SGH-T849|SGH-T859|SGH-T869|SPH-P100|GT-P3100|GT-P3108|GT-P3110|GT-P5100|GT-P5110|GT-P6200|GT-P7320|GT-P7511|GT-N8000|GT-P8510|SGH-I497|SPH-P500|SGH-T779|SCH-I705|SCH-I915|GT-N8013|GT-P3113|GT-P5113|GT-P8110|GT-N8010|GT-N8005|GT-N8020|GT-P1013|GT-P6201|GT-P7501|GT-N5100|GT-N5105|GT-N5110|SHV-E140K|SHV-E140L|SHV-E140S|SHV-E150S|SHV-E230K|SHV-E230L|SHV-E230S|SHW-M180K|SHW-M180L|SHW-M180S|SHW-M180W|SHW-M300W|SHW-M305W|SHW-M380K|SHW-M380S|SHW-M380W|SHW-M430W|SHW-M480K|SHW-M480S|SHW-M480W|SHW-M485W|SHW-M486W|SHW-M500W|GT-I9228|SCH-P739|SCH-I925|GT-I9200|GT-P5200|GT-P5210|GT-P5210X|SM-T311|SM-T310|SM-T310X|SM-T210|SM-T210R|SM-T211|SM-P600|SM-P601|SM-P605|SM-P900|SM-P901|SM-T217|SM-T217A|SM-T217S|SM-P6000|SM-T3100|SGH-I467|XE500|SM-T110|GT-P5220|GT-I9200X|GT-N5110X|GT-N5120|SM-P905|SM-T111|SM-T2105|SM-T315|SM-T320|SM-T320X|SM-T321|SM-T520|SM-T525|SM-T530NU|SM-T230NU|SM-T330NU|SM-T900|XE500T1C|SM-P605V|SM-P905V|SM-T337V|SM-T537V|SM-T707V|SM-T807V|SM-P600X|SM-P900X|SM-T210X|SM-T230|SM-T230X|SM-T325|GT-P7503|SM-T531|SM-T330|SM-T530|SM-T705|SM-T705C|SM-T535|SM-T331|SM-T800|SM-T700|SM-T537|SM-T807|SM-P907A|SM-T337A|SM-T537A|SM-T707A|SM-T807A|SM-T237|SM-T807P|SM-P607T|SM-T217T|SM-T337T|SM-T807T|SM-T116NQ|SM-T116BU|SM-P550|SM-T350|SM-T550|SM-T9000|SM-P9000|SM-T705Y|SM-T805|GT-P3113|SM-T710|SM-T810|SM-T815|SM-T360|SM-T533|SM-T113|SM-T335|SM-T715|SM-T560|SM-T670|SM-T677|SM-T377|SM-T567|SM-T357T|SM-T555|SM-T561|SM-T713|SM-T719|SM-T813|SM-T819|SM-T580|SM-T355Y?|SM-T280|SM-T817A|SM-T820|SM-W700|SM-P580|SM-T587|SM-P350|SM-P555M|SM-P355M|SM-T113NU|SM-T815Y|SM-T585|SM-T285|SM-T825|SM-W708|SM-T835|SM-T830|SM-T837V|SM-T720|SM-T510|SM-T387V|SM-P610|SM-T290|SM-T515|SM-T590|SM-T595|SM-T725|SM-T817P|SM-P585N0|SM-T395|SM-T295|SM-T865|SM-P610N|SM-P615|SM-T970|SM-T380|SM-T5950|SM-T905|SM-T231|SM-T500|SM-T860",Kindle:"Kindle|Silk.*Accelerated|Android.*\\b(KFOT|KFTT|KFJWI|KFJWA|KFOTE|KFSOWI|KFTHWI|KFTHWA|KFAPWI|KFAPWA|WFJWAE|KFSAWA|KFSAWI|KFASWI|KFARWI|KFFOWI|KFGIWI|KFMEWI)\\b|Android.*Silk/[0-9.]+ like Chrome/[0-9.]+ (?!Mobile)",SurfaceTablet:"Windows NT [0-9.]+; ARM;.*(Tablet|ARMBJS)",HPTablet:"HP Slate (7|8|10)|HP ElitePad 900|hp-tablet|EliteBook.*Touch|HP 8|Slate 21|HP SlateBook 10",AsusTablet:"^.*PadFone((?!Mobile).)*$|Transformer|TF101|TF101G|TF300T|TF300TG|TF300TL|TF700T|TF700KL|TF701T|TF810C|ME171|ME301T|ME302C|ME371MG|ME370T|ME372MG|ME172V|ME173X|ME400C|Slider SL101|\\bK00F\\b|\\bK00C\\b|\\bK00E\\b|\\bK00L\\b|TX201LA|ME176C|ME102A|\\bM80TA\\b|ME372CL|ME560CG|ME372CG|ME302KL| K010 | K011 | K017 | K01E |ME572C|ME103K|ME170C|ME171C|\\bME70C\\b|ME581C|ME581CL|ME8510C|ME181C|P01Y|PO1MA|P01Z|\\bP027\\b|\\bP024\\b|\\bP00C\\b",BlackBerryTablet:"PlayBook|RIM Tablet",HTCtablet:"HTC_Flyer_P512|HTC Flyer|HTC Jetstream|HTC-P715a|HTC EVO View 4G|PG41200|PG09410",MotorolaTablet:"xoom|sholest|MZ615|MZ605|MZ505|MZ601|MZ602|MZ603|MZ604|MZ606|MZ607|MZ608|MZ609|MZ615|MZ616|MZ617",NookTablet:"Android.*Nook|NookColor|nook browser|BNRV200|BNRV200A|BNTV250|BNTV250A|BNTV400|BNTV600|LogicPD Zoom2",AcerTablet:"Android.*; \\b(A100|A101|A110|A200|A210|A211|A500|A501|A510|A511|A700|A701|W500|W500P|W501|W501P|W510|W511|W700|G100|G100W|B1-A71|B1-710|B1-711|A1-810|A1-811|A1-830)\\b|W3-810|\\bA3-A10\\b|\\bA3-A11\\b|\\bA3-A20\\b|\\bA3-A30|A3-A40",ToshibaTablet:"Android.*(AT100|AT105|AT200|AT205|AT270|AT275|AT300|AT305|AT1S5|AT500|AT570|AT700|AT830)|TOSHIBA.*FOLIO",LGTablet:"\\bL-06C|LG-V909|LG-V900|LG-V700|LG-V510|LG-V500|LG-V410|LG-V400|LG-VK810\\b",FujitsuTablet:"Android.*\\b(F-01D|F-02F|F-05E|F-10D|M532|Q572)\\b",PrestigioTablet:"PMP3170B|PMP3270B|PMP3470B|PMP7170B|PMP3370B|PMP3570C|PMP5870C|PMP3670B|PMP5570C|PMP5770D|PMP3970B|PMP3870C|PMP5580C|PMP5880D|PMP5780D|PMP5588C|PMP7280C|PMP7280C3G|PMP7280|PMP7880D|PMP5597D|PMP5597|PMP7100D|PER3464|PER3274|PER3574|PER3884|PER5274|PER5474|PMP5097CPRO|PMP5097|PMP7380D|PMP5297C|PMP5297C_QUAD|PMP812E|PMP812E3G|PMP812F|PMP810E|PMP880TD|PMT3017|PMT3037|PMT3047|PMT3057|PMT7008|PMT5887|PMT5001|PMT5002",LenovoTablet:"Lenovo TAB|Idea(Tab|Pad)( A1|A10| K1|)|ThinkPad([ ]+)?Tablet|YT3-850M|YT3-X90L|YT3-X90F|YT3-X90X|Lenovo.*(S2109|S2110|S5000|S6000|K3011|A3000|A3500|A1000|A2107|A2109|A1107|A5500|A7600|B6000|B8000|B8080)(-|)(FL|F|HV|H|)|TB-X103F|TB-X304X|TB-X304F|TB-X304L|TB-X505F|TB-X505L|TB-X505X|TB-X605F|TB-X605L|TB-8703F|TB-8703X|TB-8703N|TB-8704N|TB-8704F|TB-8704X|TB-8704V|TB-7304F|TB-7304I|TB-7304X|Tab2A7-10F|Tab2A7-20F|TB2-X30L|YT3-X50L|YT3-X50F|YT3-X50M|YT-X705F|YT-X703F|YT-X703L|YT-X705L|YT-X705X|TB2-X30F|TB2-X30L|TB2-X30M|A2107A-F|A2107A-H|TB3-730F|TB3-730M|TB3-730X|TB-7504F|TB-7504X|TB-X704F|TB-X104F|TB3-X70F|TB-X705F|TB-8504F|TB3-X70L|TB3-710F|TB-X704L",DellTablet:"Venue 11|Venue 8|Venue 7|Dell Streak 10|Dell Streak 7",YarvikTablet:"Android.*\\b(TAB210|TAB211|TAB224|TAB250|TAB260|TAB264|TAB310|TAB360|TAB364|TAB410|TAB411|TAB420|TAB424|TAB450|TAB460|TAB461|TAB464|TAB465|TAB467|TAB468|TAB07-100|TAB07-101|TAB07-150|TAB07-151|TAB07-152|TAB07-200|TAB07-201-3G|TAB07-210|TAB07-211|TAB07-212|TAB07-214|TAB07-220|TAB07-400|TAB07-485|TAB08-150|TAB08-200|TAB08-201-3G|TAB08-201-30|TAB09-100|TAB09-211|TAB09-410|TAB10-150|TAB10-201|TAB10-211|TAB10-400|TAB10-410|TAB13-201|TAB274EUK|TAB275EUK|TAB374EUK|TAB462EUK|TAB474EUK|TAB9-200)\\b",MedionTablet:"Android.*\\bOYO\\b|LIFE.*(P9212|P9514|P9516|S9512)|LIFETAB",ArnovaTablet:"97G4|AN10G2|AN7bG3|AN7fG3|AN8G3|AN8cG3|AN7G3|AN9G3|AN7dG3|AN7dG3ST|AN7dG3ChildPad|AN10bG3|AN10bG3DT|AN9G2",IntensoTablet:"INM8002KP|INM1010FP|INM805ND|Intenso Tab|TAB1004",IRUTablet:"M702pro",MegafonTablet:"MegaFon V9|\\bZTE V9\\b|Android.*\\bMT7A\\b",EbodaTablet:"E-Boda (Supreme|Impresspeed|Izzycomm|Essential)",AllViewTablet:"Allview.*(Viva|Alldro|City|Speed|All TV|Frenzy|Quasar|Shine|TX1|AX1|AX2)",ArchosTablet:"\\b(101G9|80G9|A101IT)\\b|Qilive 97R|Archos5|\\bARCHOS (70|79|80|90|97|101|FAMILYPAD|)(b|c|)(G10| Cobalt| TITANIUM(HD|)| Xenon| Neon|XSK| 2| XS 2| PLATINUM| CARBON|GAMEPAD)\\b",AinolTablet:"NOVO7|NOVO8|NOVO10|Novo7Aurora|Novo7Basic|NOVO7PALADIN|novo9-Spark",NokiaLumiaTablet:"Lumia 2520",SonyTablet:"Sony.*Tablet|Xperia Tablet|Sony Tablet S|SO-03E|SGPT12|SGPT13|SGPT114|SGPT121|SGPT122|SGPT123|SGPT111|SGPT112|SGPT113|SGPT131|SGPT132|SGPT133|SGPT211|SGPT212|SGPT213|SGP311|SGP312|SGP321|EBRD1101|EBRD1102|EBRD1201|SGP351|SGP341|SGP511|SGP512|SGP521|SGP541|SGP551|SGP621|SGP641|SGP612|SOT31|SGP771|SGP611|SGP612|SGP712",PhilipsTablet:"\\b(PI2010|PI3000|PI3100|PI3105|PI3110|PI3205|PI3210|PI3900|PI4010|PI7000|PI7100)\\b",CubeTablet:"Android.*(K8GT|U9GT|U10GT|U16GT|U17GT|U18GT|U19GT|U20GT|U23GT|U30GT)|CUBE U8GT",CobyTablet:"MID1042|MID1045|MID1125|MID1126|MID7012|MID7014|MID7015|MID7034|MID7035|MID7036|MID7042|MID7048|MID7127|MID8042|MID8048|MID8127|MID9042|MID9740|MID9742|MID7022|MID7010",MIDTablet:"M9701|M9000|M9100|M806|M1052|M806|T703|MID701|MID713|MID710|MID727|MID760|MID830|MID728|MID933|MID125|MID810|MID732|MID120|MID930|MID800|MID731|MID900|MID100|MID820|MID735|MID980|MID130|MID833|MID737|MID960|MID135|MID860|MID736|MID140|MID930|MID835|MID733|MID4X10",MSITablet:"MSI \\b(Primo 73K|Primo 73L|Primo 81L|Primo 77|Primo 93|Primo 75|Primo 76|Primo 73|Primo 81|Primo 91|Primo 90|Enjoy 71|Enjoy 7|Enjoy 10)\\b",SMiTTablet:"Android.*(\\bMID\\b|MID-560|MTV-T1200|MTV-PND531|MTV-P1101|MTV-PND530)",RockChipTablet:"Android.*(RK2818|RK2808A|RK2918|RK3066)|RK2738|RK2808A",FlyTablet:"IQ310|Fly Vision",bqTablet:"Android.*(bq)?.*\\b(Elcano|Curie|Edison|Maxwell|Kepler|Pascal|Tesla|Hypatia|Platon|Newton|Livingstone|Cervantes|Avant|Aquaris ([E|M]10|M8))\\b|Maxwell.*Lite|Maxwell.*Plus",HuaweiTablet:"MediaPad|MediaPad 7 Youth|IDEOS S7|S7-201c|S7-202u|S7-101|S7-103|S7-104|S7-105|S7-106|S7-201|S7-Slim|M2-A01L|BAH-L09|BAH-W09|AGS-L09|CMR-AL19",NecTablet:"\\bN-06D|\\bN-08D",PantechTablet:"Pantech.*P4100",BronchoTablet:"Broncho.*(N701|N708|N802|a710)",VersusTablet:"TOUCHPAD.*[78910]|\\bTOUCHTAB\\b",ZyncTablet:"z1000|Z99 2G|z930|z990|z909|Z919|z900",PositivoTablet:"TB07STA|TB10STA|TB07FTA|TB10FTA",NabiTablet:"Android.*\\bNabi",KoboTablet:"Kobo Touch|\\bK080\\b|\\bVox\\b Build|\\bArc\\b Build",DanewTablet:"DSlide.*\\b(700|701R|702|703R|704|802|970|971|972|973|974|1010|1012)\\b",TexetTablet:"NaviPad|TB-772A|TM-7045|TM-7055|TM-9750|TM-7016|TM-7024|TM-7026|TM-7041|TM-7043|TM-7047|TM-8041|TM-9741|TM-9747|TM-9748|TM-9751|TM-7022|TM-7021|TM-7020|TM-7011|TM-7010|TM-7023|TM-7025|TM-7037W|TM-7038W|TM-7027W|TM-9720|TM-9725|TM-9737W|TM-1020|TM-9738W|TM-9740|TM-9743W|TB-807A|TB-771A|TB-727A|TB-725A|TB-719A|TB-823A|TB-805A|TB-723A|TB-715A|TB-707A|TB-705A|TB-709A|TB-711A|TB-890HD|TB-880HD|TB-790HD|TB-780HD|TB-770HD|TB-721HD|TB-710HD|TB-434HD|TB-860HD|TB-840HD|TB-760HD|TB-750HD|TB-740HD|TB-730HD|TB-722HD|TB-720HD|TB-700HD|TB-500HD|TB-470HD|TB-431HD|TB-430HD|TB-506|TB-504|TB-446|TB-436|TB-416|TB-146SE|TB-126SE",PlaystationTablet:"Playstation.*(Portable|Vita)",TrekstorTablet:"ST10416-1|VT10416-1|ST70408-1|ST702xx-1|ST702xx-2|ST80208|ST97216|ST70104-2|VT10416-2|ST10216-2A|SurfTab",PyleAudioTablet:"\\b(PTBL10CEU|PTBL10C|PTBL72BC|PTBL72BCEU|PTBL7CEU|PTBL7C|PTBL92BC|PTBL92BCEU|PTBL9CEU|PTBL9CUK|PTBL9C)\\b",AdvanTablet:"Android.* \\b(E3A|T3X|T5C|T5B|T3E|T3C|T3B|T1J|T1F|T2A|T1H|T1i|E1C|T1-E|T5-A|T4|E1-B|T2Ci|T1-B|T1-D|O1-A|E1-A|T1-A|T3A|T4i)\\b ",DanyTechTablet:"Genius Tab G3|Genius Tab S2|Genius Tab Q3|Genius Tab G4|Genius Tab Q4|Genius Tab G-II|Genius TAB GII|Genius TAB GIII|Genius Tab S1",GalapadTablet:"Android [0-9.]+; [a-z-]+; \\bG1\\b",MicromaxTablet:"Funbook|Micromax.*\\b(P250|P560|P360|P362|P600|P300|P350|P500|P275)\\b",KarbonnTablet:"Android.*\\b(A39|A37|A34|ST8|ST10|ST7|Smart Tab3|Smart Tab2)\\b",AllFineTablet:"Fine7 Genius|Fine7 Shine|Fine7 Air|Fine8 Style|Fine9 More|Fine10 Joy|Fine11 Wide",PROSCANTablet:"\\b(PEM63|PLT1023G|PLT1041|PLT1044|PLT1044G|PLT1091|PLT4311|PLT4311PL|PLT4315|PLT7030|PLT7033|PLT7033D|PLT7035|PLT7035D|PLT7044K|PLT7045K|PLT7045KB|PLT7071KG|PLT7072|PLT7223G|PLT7225G|PLT7777G|PLT7810K|PLT7849G|PLT7851G|PLT7852G|PLT8015|PLT8031|PLT8034|PLT8036|PLT8080K|PLT8082|PLT8088|PLT8223G|PLT8234G|PLT8235G|PLT8816K|PLT9011|PLT9045K|PLT9233G|PLT9735|PLT9760G|PLT9770G)\\b",YONESTablet:"BQ1078|BC1003|BC1077|RK9702|BC9730|BC9001|IT9001|BC7008|BC7010|BC708|BC728|BC7012|BC7030|BC7027|BC7026",ChangJiaTablet:"TPC7102|TPC7103|TPC7105|TPC7106|TPC7107|TPC7201|TPC7203|TPC7205|TPC7210|TPC7708|TPC7709|TPC7712|TPC7110|TPC8101|TPC8103|TPC8105|TPC8106|TPC8203|TPC8205|TPC8503|TPC9106|TPC9701|TPC97101|TPC97103|TPC97105|TPC97106|TPC97111|TPC97113|TPC97203|TPC97603|TPC97809|TPC97205|TPC10101|TPC10103|TPC10106|TPC10111|TPC10203|TPC10205|TPC10503",GUTablet:"TX-A1301|TX-M9002|Q702|kf026",PointOfViewTablet:"TAB-P506|TAB-navi-7-3G-M|TAB-P517|TAB-P-527|TAB-P701|TAB-P703|TAB-P721|TAB-P731N|TAB-P741|TAB-P825|TAB-P905|TAB-P925|TAB-PR945|TAB-PL1015|TAB-P1025|TAB-PI1045|TAB-P1325|TAB-PROTAB[0-9]+|TAB-PROTAB25|TAB-PROTAB26|TAB-PROTAB27|TAB-PROTAB26XL|TAB-PROTAB2-IPS9|TAB-PROTAB30-IPS9|TAB-PROTAB25XXL|TAB-PROTAB26-IPS10|TAB-PROTAB30-IPS10",OvermaxTablet:"OV-(SteelCore|NewBase|Basecore|Baseone|Exellen|Quattor|EduTab|Solution|ACTION|BasicTab|TeddyTab|MagicTab|Stream|TB-08|TB-09)|Qualcore 1027",HCLTablet:"HCL.*Tablet|Connect-3G-2.0|Connect-2G-2.0|ME Tablet U1|ME Tablet U2|ME Tablet G1|ME Tablet X1|ME Tablet Y2|ME Tablet Sync",DPSTablet:"DPS Dream 9|DPS Dual 7",VistureTablet:"V97 HD|i75 3G|Visture V4( HD)?|Visture V5( HD)?|Visture V10",CrestaTablet:"CTP(-)?810|CTP(-)?818|CTP(-)?828|CTP(-)?838|CTP(-)?888|CTP(-)?978|CTP(-)?980|CTP(-)?987|CTP(-)?988|CTP(-)?989",MediatekTablet:"\\bMT8125|MT8389|MT8135|MT8377\\b",ConcordeTablet:"Concorde([ ]+)?Tab|ConCorde ReadMan",GoCleverTablet:"GOCLEVER TAB|A7GOCLEVER|M1042|M7841|M742|R1042BK|R1041|TAB A975|TAB A7842|TAB A741|TAB A741L|TAB M723G|TAB M721|TAB A1021|TAB I921|TAB R721|TAB I720|TAB T76|TAB R70|TAB R76.2|TAB R106|TAB R83.2|TAB M813G|TAB I721|GCTA722|TAB I70|TAB I71|TAB S73|TAB R73|TAB R74|TAB R93|TAB R75|TAB R76.1|TAB A73|TAB A93|TAB A93.2|TAB T72|TAB R83|TAB R974|TAB R973|TAB A101|TAB A103|TAB A104|TAB A104.2|R105BK|M713G|A972BK|TAB A971|TAB R974.2|TAB R104|TAB R83.3|TAB A1042",ModecomTablet:"FreeTAB 9000|FreeTAB 7.4|FreeTAB 7004|FreeTAB 7800|FreeTAB 2096|FreeTAB 7.5|FreeTAB 1014|FreeTAB 1001 |FreeTAB 8001|FreeTAB 9706|FreeTAB 9702|FreeTAB 7003|FreeTAB 7002|FreeTAB 1002|FreeTAB 7801|FreeTAB 1331|FreeTAB 1004|FreeTAB 8002|FreeTAB 8014|FreeTAB 9704|FreeTAB 1003",VoninoTablet:"\\b(Argus[ _]?S|Diamond[ _]?79HD|Emerald[ _]?78E|Luna[ _]?70C|Onyx[ _]?S|Onyx[ _]?Z|Orin[ _]?HD|Orin[ _]?S|Otis[ _]?S|SpeedStar[ _]?S|Magnet[ _]?M9|Primus[ _]?94[ _]?3G|Primus[ _]?94HD|Primus[ _]?QS|Android.*\\bQ8\\b|Sirius[ _]?EVO[ _]?QS|Sirius[ _]?QS|Spirit[ _]?S)\\b",ECSTablet:"V07OT2|TM105A|S10OT1|TR10CS1",StorexTablet:"eZee[_']?(Tab|Go)[0-9]+|TabLC7|Looney Tunes Tab",VodafoneTablet:"SmartTab([ ]+)?[0-9]+|SmartTabII10|SmartTabII7|VF-1497|VFD 1400",EssentielBTablet:"Smart[ ']?TAB[ ]+?[0-9]+|Family[ ']?TAB2",RossMoorTablet:"RM-790|RM-997|RMD-878G|RMD-974R|RMT-705A|RMT-701|RME-601|RMT-501|RMT-711",iMobileTablet:"i-mobile i-note",TolinoTablet:"tolino tab [0-9.]+|tolino shine",AudioSonicTablet:"\\bC-22Q|T7-QC|T-17B|T-17P\\b",AMPETablet:"Android.* A78 ",SkkTablet:"Android.* (SKYPAD|PHOENIX|CYCLOPS)",TecnoTablet:"TECNO P9|TECNO DP8D",JXDTablet:"Android.* \\b(F3000|A3300|JXD5000|JXD3000|JXD2000|JXD300B|JXD300|S5800|S7800|S602b|S5110b|S7300|S5300|S602|S603|S5100|S5110|S601|S7100a|P3000F|P3000s|P101|P200s|P1000m|P200m|P9100|P1000s|S6600b|S908|P1000|P300|S18|S6600|S9100)\\b",iJoyTablet:"Tablet (Spirit 7|Essentia|Galatea|Fusion|Onix 7|Landa|Titan|Scooby|Deox|Stella|Themis|Argon|Unique 7|Sygnus|Hexen|Finity 7|Cream|Cream X2|Jade|Neon 7|Neron 7|Kandy|Scape|Saphyr 7|Rebel|Biox|Rebel|Rebel 8GB|Myst|Draco 7|Myst|Tab7-004|Myst|Tadeo Jones|Tablet Boing|Arrow|Draco Dual Cam|Aurix|Mint|Amity|Revolution|Finity 9|Neon 9|T9w|Amity 4GB Dual Cam|Stone 4GB|Stone 8GB|Andromeda|Silken|X2|Andromeda II|Halley|Flame|Saphyr 9,7|Touch 8|Planet|Triton|Unique 10|Hexen 10|Memphis 4GB|Memphis 8GB|Onix 10)",FX2Tablet:"FX2 PAD7|FX2 PAD10",XoroTablet:"KidsPAD 701|PAD[ ]?712|PAD[ ]?714|PAD[ ]?716|PAD[ ]?717|PAD[ ]?718|PAD[ ]?720|PAD[ ]?721|PAD[ ]?722|PAD[ ]?790|PAD[ ]?792|PAD[ ]?900|PAD[ ]?9715D|PAD[ ]?9716DR|PAD[ ]?9718DR|PAD[ ]?9719QR|PAD[ ]?9720QR|TelePAD1030|Telepad1032|TelePAD730|TelePAD731|TelePAD732|TelePAD735Q|TelePAD830|TelePAD9730|TelePAD795|MegaPAD 1331|MegaPAD 1851|MegaPAD 2151",ViewsonicTablet:"ViewPad 10pi|ViewPad 10e|ViewPad 10s|ViewPad E72|ViewPad7|ViewPad E100|ViewPad 7e|ViewSonic VB733|VB100a",VerizonTablet:"QTAQZ3|QTAIR7|QTAQTZ3|QTASUN1|QTASUN2|QTAXIA1",OdysTablet:"LOOX|XENO10|ODYS[ -](Space|EVO|Xpress|NOON)|\\bXELIO\\b|Xelio10Pro|XELIO7PHONETAB|XELIO10EXTREME|XELIOPT2|NEO_QUAD10",CaptivaTablet:"CAPTIVA PAD",IconbitTablet:"NetTAB|NT-3702|NT-3702S|NT-3702S|NT-3603P|NT-3603P|NT-0704S|NT-0704S|NT-3805C|NT-3805C|NT-0806C|NT-0806C|NT-0909T|NT-0909T|NT-0907S|NT-0907S|NT-0902S|NT-0902S",TeclastTablet:"T98 4G|\\bP80\\b|\\bX90HD\\b|X98 Air|X98 Air 3G|\\bX89\\b|P80 3G|\\bX80h\\b|P98 Air|\\bX89HD\\b|P98 3G|\\bP90HD\\b|P89 3G|X98 3G|\\bP70h\\b|P79HD 3G|G18d 3G|\\bP79HD\\b|\\bP89s\\b|\\bA88\\b|\\bP10HD\\b|\\bP19HD\\b|G18 3G|\\bP78HD\\b|\\bA78\\b|\\bP75\\b|G17s 3G|G17h 3G|\\bP85t\\b|\\bP90\\b|\\bP11\\b|\\bP98t\\b|\\bP98HD\\b|\\bG18d\\b|\\bP85s\\b|\\bP11HD\\b|\\bP88s\\b|\\bA80HD\\b|\\bA80se\\b|\\bA10h\\b|\\bP89\\b|\\bP78s\\b|\\bG18\\b|\\bP85\\b|\\bA70h\\b|\\bA70\\b|\\bG17\\b|\\bP18\\b|\\bA80s\\b|\\bA11s\\b|\\bP88HD\\b|\\bA80h\\b|\\bP76s\\b|\\bP76h\\b|\\bP98\\b|\\bA10HD\\b|\\bP78\\b|\\bP88\\b|\\bA11\\b|\\bA10t\\b|\\bP76a\\b|\\bP76t\\b|\\bP76e\\b|\\bP85HD\\b|\\bP85a\\b|\\bP86\\b|\\bP75HD\\b|\\bP76v\\b|\\bA12\\b|\\bP75a\\b|\\bA15\\b|\\bP76Ti\\b|\\bP81HD\\b|\\bA10\\b|\\bT760VE\\b|\\bT720HD\\b|\\bP76\\b|\\bP73\\b|\\bP71\\b|\\bP72\\b|\\bT720SE\\b|\\bC520Ti\\b|\\bT760\\b|\\bT720VE\\b|T720-3GE|T720-WiFi",OndaTablet:"\\b(V975i|Vi30|VX530|V701|Vi60|V701s|Vi50|V801s|V719|Vx610w|VX610W|V819i|Vi10|VX580W|Vi10|V711s|V813|V811|V820w|V820|Vi20|V711|VI30W|V712|V891w|V972|V819w|V820w|Vi60|V820w|V711|V813s|V801|V819|V975s|V801|V819|V819|V818|V811|V712|V975m|V101w|V961w|V812|V818|V971|V971s|V919|V989|V116w|V102w|V973|Vi40)\\b[\\s]+|V10 \\b4G\\b",JaytechTablet:"TPC-PA762",BlaupunktTablet:"Endeavour 800NG|Endeavour 1010",DigmaTablet:"\\b(iDx10|iDx9|iDx8|iDx7|iDxD7|iDxD8|iDsQ8|iDsQ7|iDsQ8|iDsD10|iDnD7|3TS804H|iDsQ11|iDj7|iDs10)\\b",EvolioTablet:"ARIA_Mini_wifi|Aria[ _]Mini|Evolio X10|Evolio X7|Evolio X8|\\bEvotab\\b|\\bNeura\\b",LavaTablet:"QPAD E704|\\bIvoryS\\b|E-TAB IVORY|\\bE-TAB\\b",AocTablet:"MW0811|MW0812|MW0922|MTK8382|MW1031|MW0831|MW0821|MW0931|MW0712",MpmanTablet:"MP11 OCTA|MP10 OCTA|MPQC1114|MPQC1004|MPQC994|MPQC974|MPQC973|MPQC804|MPQC784|MPQC780|\\bMPG7\\b|MPDCG75|MPDCG71|MPDC1006|MP101DC|MPDC9000|MPDC905|MPDC706HD|MPDC706|MPDC705|MPDC110|MPDC100|MPDC99|MPDC97|MPDC88|MPDC8|MPDC77|MP709|MID701|MID711|MID170|MPDC703|MPQC1010",CelkonTablet:"CT695|CT888|CT[\\s]?910|CT7 Tab|CT9 Tab|CT3 Tab|CT2 Tab|CT1 Tab|C820|C720|\\bCT-1\\b",WolderTablet:"miTab \\b(DIAMOND|SPACE|BROOKLYN|NEO|FLY|MANHATTAN|FUNK|EVOLUTION|SKY|GOCAR|IRON|GENIUS|POP|MINT|EPSILON|BROADWAY|JUMP|HOP|LEGEND|NEW AGE|LINE|ADVANCE|FEEL|FOLLOW|LIKE|LINK|LIVE|THINK|FREEDOM|CHICAGO|CLEVELAND|BALTIMORE-GH|IOWA|BOSTON|SEATTLE|PHOENIX|DALLAS|IN 101|MasterChef)\\b",MediacomTablet:"M-MPI10C3G|M-SP10EG|M-SP10EGP|M-SP10HXAH|M-SP7HXAH|M-SP10HXBH|M-SP8HXAH|M-SP8MXA",MiTablet:"\\bMI PAD\\b|\\bHM NOTE 1W\\b",NibiruTablet:"Nibiru M1|Nibiru Jupiter One",NexoTablet:"NEXO NOVA|NEXO 10|NEXO AVIO|NEXO FREE|NEXO GO|NEXO EVO|NEXO 3G|NEXO SMART|NEXO KIDDO|NEXO MOBI",LeaderTablet:"TBLT10Q|TBLT10I|TBL-10WDKB|TBL-10WDKBO2013|TBL-W230V2|TBL-W450|TBL-W500|SV572|TBLT7I|TBA-AC7-8G|TBLT79|TBL-8W16|TBL-10W32|TBL-10WKB|TBL-W100",UbislateTablet:"UbiSlate[\\s]?7C",PocketBookTablet:"Pocketbook",KocasoTablet:"\\b(TB-1207)\\b",HisenseTablet:"\\b(F5281|E2371)\\b",Hudl:"Hudl HT7S3|Hudl 2",TelstraTablet:"T-Hub2",GenericTablet:"Android.*\\b97D\\b|Tablet(?!.*PC)|BNTV250A|MID-WCDMA|LogicPD Zoom2|\\bA7EB\\b|CatNova8|A1_07|CT704|CT1002|\\bM721\\b|rk30sdk|\\bEVOTAB\\b|M758A|ET904|ALUMIUM10|Smartfren Tab|Endeavour 1010|Tablet-PC-4|Tagi Tab|\\bM6pro\\b|CT1020W|arc 10HD|\\bTP750\\b|\\bQTAQZ3\\b|WVT101|TM1088|KT107"},oss:{AndroidOS:"Android",BlackBerryOS:"blackberry|\\bBB10\\b|rim tablet os",PalmOS:"PalmOS|avantgo|blazer|elaine|hiptop|palm|plucker|xiino",SymbianOS:"Symbian|SymbOS|Series60|Series40|SYB-[0-9]+|\\bS60\\b",WindowsMobileOS:"Windows CE.*(PPC|Smartphone|Mobile|[0-9]{3}x[0-9]{3})|Windows Mobile|Windows Phone [0-9.]+|WCE;",WindowsPhoneOS:"Windows Phone 10.0|Windows Phone 8.1|Windows Phone 8.0|Windows Phone OS|XBLWP7|ZuneWP7|Windows NT 6.[23]; ARM;",iOS:"\\biPhone.*Mobile|\\biPod|\\biPad|AppleCoreMedia",iPadOS:"CPU OS 13",SailfishOS:"Sailfish",MeeGoOS:"MeeGo",MaemoOS:"Maemo",JavaOS:"J2ME/|\\bMIDP\\b|\\bCLDC\\b",webOS:"webOS|hpwOS",badaOS:"\\bBada\\b",BREWOS:"BREW"},uas:{Chrome:"\\bCrMo\\b|CriOS|Android.*Chrome/[.0-9]* (Mobile)?",Dolfin:"\\bDolfin\\b",Opera:"Opera.*Mini|Opera.*Mobi|Android.*Opera|Mobile.*OPR/[0-9.]+$|Coast/[0-9.]+",Skyfire:"Skyfire",Edge:"\\bEdgiOS\\b|Mobile Safari/[.0-9]* Edge",IE:"IEMobile|MSIEMobile",Firefox:"fennec|firefox.*maemo|(Mobile|Tablet).*Firefox|Firefox.*Mobile|FxiOS",Bolt:"bolt",TeaShark:"teashark",Blazer:"Blazer",Safari:"Version((?!\\bEdgiOS\\b).)*Mobile.*Safari|Safari.*Mobile|MobileSafari",WeChat:"\\bMicroMessenger\\b",UCBrowser:"UC.*Browser|UCWEB",baiduboxapp:"baiduboxapp",baidubrowser:"baidubrowser",DiigoBrowser:"DiigoBrowser",Mercury:"\\bMercury\\b",ObigoBrowser:"Obigo",NetFront:"NF-Browser",GenericBrowser:"NokiaBrowser|OviBrowser|OneBrowser|TwonkyBeamBrowser|SEMC.*Browser|FlyFlow|Minimo|NetFront|Novarra-Vision|MQQBrowser|MicroMessenger",PaleMoon:"Android.*PaleMoon|Mobile.*PaleMoon"},props:{Mobile:"Mobile/[VER]",Build:"Build/[VER]",Version:"Version/[VER]",VendorID:"VendorID/[VER]",iPad:"iPad.*CPU[a-z ]+[VER]",iPhone:"iPhone.*CPU[a-z ]+[VER]",iPod:"iPod.*CPU[a-z ]+[VER]",Kindle:"Kindle/[VER]",Chrome:["Chrome/[VER]","CriOS/[VER]","CrMo/[VER]"],Coast:["Coast/[VER]"],Dolfin:"Dolfin/[VER]",Firefox:["Firefox/[VER]","FxiOS/[VER]"],Fennec:"Fennec/[VER]",Edge:"Edge/[VER]",IE:["IEMobile/[VER];","IEMobile [VER]","MSIE [VER];","Trident/[0-9.]+;.*rv:[VER]"],NetFront:"NetFront/[VER]",NokiaBrowser:"NokiaBrowser/[VER]",Opera:[" OPR/[VER]","Opera Mini/[VER]","Version/[VER]"],"Opera Mini":"Opera Mini/[VER]","Opera Mobi":"Version/[VER]",UCBrowser:["UCWEB[VER]","UC.*Browser/[VER]"],MQQBrowser:"MQQBrowser/[VER]",MicroMessenger:"MicroMessenger/[VER]",baiduboxapp:"baiduboxapp/[VER]",baidubrowser:"baidubrowser/[VER]",SamsungBrowser:"SamsungBrowser/[VER]",Iron:"Iron/[VER]",Safari:["Version/[VER]","Safari/[VER]"],Skyfire:"Skyfire/[VER]",Tizen:"Tizen/[VER]",Webkit:"webkit[ /][VER]",PaleMoon:"PaleMoon/[VER]",SailfishBrowser:"SailfishBrowser/[VER]",Gecko:"Gecko/[VER]",Trident:"Trident/[VER]",Presto:"Presto/[VER]",Goanna:"Goanna/[VER]",iOS:" \\bi?OS\\b [VER][ ;]{1}",Android:"Android [VER]",Sailfish:"Sailfish [VER]",BlackBerry:["BlackBerry[\\w]+/[VER]","BlackBerry.*Version/[VER]","Version/[VER]"],BREW:"BREW [VER]",Java:"Java/[VER]","Windows Phone OS":["Windows Phone OS [VER]","Windows Phone [VER]"],"Windows Phone":"Windows Phone [VER]","Windows CE":"Windows CE/[VER]","Windows NT":"Windows NT [VER]",Symbian:["SymbianOS/[VER]","Symbian/[VER]"],webOS:["webOS/[VER]","hpwOS/[VER];"]},utils:{Bot:"Googlebot|facebookexternalhit|Google-AMPHTML|s~amp-validator|AdsBot-Google|Google Keyword Suggestion|Facebot|YandexBot|YandexMobileBot|bingbot|ia_archiver|AhrefsBot|Ezooms|GSLFbot|WBSearchBot|Twitterbot|TweetmemeBot|Twikle|PaperLiBot|Wotbox|UnwindFetchor|Exabot|MJ12bot|YandexImages|TurnitinBot|Pingdom|contentkingapp|AspiegelBot",MobileBot:"Googlebot-Mobile|AdsBot-Google-Mobile|YahooSeeker/M1A1-R2D2",DesktopMode:"WPDesktop",TV:"SonyDTV|HbbTV",WebKit:"(webkit)[ /]([\\w.]+)",Console:"\\b(Nintendo|Nintendo WiiU|Nintendo 3DS|Nintendo Switch|PLAYSTATION|Xbox)\\b",Watch:"SM-V700"}},g.detectMobileBrowsers={fullPattern:/(android|bb\d+|meego).+mobile|avantgo|bada\/|blackberry|blazer|compal|elaine|fennec|hiptop|iemobile|ip(hone|od)|iris|kindle|lge |maemo|midp|mmp|mobile.+firefox|netfront|opera m(ob|in)i|palm( os)?|phone|p(ixi|re)\/|plucker|pocket|psp|series(4|6)0|symbian|treo|up\.(browser|link)|vodafone|wap|windows ce|xda|xiino/i, shortPattern:/1207|6310|6590|3gso|4thp|50[1-6]i|770s|802s|a wa|abac|ac(er|oo|s\-)|ai(ko|rn)|al(av|ca|co)|amoi|an(ex|ny|yw)|aptu|ar(ch|go)|as(te|us)|attw|au(di|\-m|r |s )|avan|be(ck|ll|nq)|bi(lb|rd)|bl(ac|az)|br(e|v)w|bumb|bw\-(n|u)|c55\/|capi|ccwa|cdm\-|cell|chtm|cldc|cmd\-|co(mp|nd)|craw|da(it|ll|ng)|dbte|dc\-s|devi|dica|dmob|do(c|p)o|ds(12|\-d)|el(49|ai)|em(l2|ul)|er(ic|k0)|esl8|ez([4-7]0|os|wa|ze)|fetc|fly(\-|_)|g1 u|g560|gene|gf\-5|g\-mo|go(\.w|od)|gr(ad|un)|haie|hcit|hd\-(m|p|t)|hei\-|hi(pt|ta)|hp( i|ip)|hs\-c|ht(c(\-| |_|a|g|p|s|t)|tp)|hu(aw|tc)|i\-(20|go|ma)|i230|iac( |\-|\/)|ibro|idea|ig01|ikom|im1k|inno|ipaq|iris|ja(t|v)a|jbro|jemu|jigs|kddi|keji|kgt( |\/)|klon|kpt |kwc\-|kyo(c|k)|le(no|xi)|lg( g|\/(k|l|u)|50|54|\-[a-w])|libw|lynx|m1\-w|m3ga|m50\/|ma(te|ui|xo)|mc(01|21|ca)|m\-cr|me(rc|ri)|mi(o8|oa|ts)|mmef|mo(01|02|bi|de|do|t(\-| |o|v)|zz)|mt(50|p1|v )|mwbp|mywa|n10[0-2]|n20[2-3]|n30(0|2)|n50(0|2|5)|n7(0(0|1)|10)|ne((c|m)\-|on|tf|wf|wg|wt)|nok(6|i)|nzph|o2im|op(ti|wv)|oran|owg1|p800|pan(a|d|t)|pdxg|pg(13|\-([1-8]|c))|phil|pire|pl(ay|uc)|pn\-2|po(ck|rt|se)|prox|psio|pt\-g|qa\-a|qc(07|12|21|32|60|\-[2-7]|i\-)|qtek|r380|r600|raks|rim9|ro(ve|zo)|s55\/|sa(ge|ma|mm|ms|ny|va)|sc(01|h\-|oo|p\-)|sdk\/|se(c(\-|0|1)|47|mc|nd|ri)|sgh\-|shar|sie(\-|m)|sk\-0|sl(45|id)|sm(al|ar|b3|it|t5)|so(ft|ny)|sp(01|h\-|v\-|v )|sy(01|mb)|t2(18|50)|t6(00|10|18)|ta(gt|lk)|tcl\-|tdg\-|tel(i|m)|tim\-|t\-mo|to(pl|sh)|ts(70|m\-|m3|m5)|tx\-9|up(\.b|g1|si)|utst|v400|v750|veri|vi(rg|te)|vk(40|5[0-3]|\-v)|vm40|voda|vulc|vx(52|53|60|61|70|80|81|83|85|98)|w3c(\-| )|webc|whit|wi(g |nc|nw)|wmlb|wonu|x700|yas\-|your|zeto|zte\-/i,tabletPattern:/android|ipad|playbook|silk/i};var h,i=Object.prototype.hasOwnProperty;return g.FALLBACK_PHONE="UnknownPhone",g.FALLBACK_TABLET="UnknownTablet",g.FALLBACK_MOBILE="UnknownMobile",h="isArray"in Array?Array.isArray:function(a){return"[object Array]"===Object.prototype.toString.call(a)},function(){var a,b,c,e,f,j,k=g.mobileDetectRules;for(a in k.props)if(i.call(k.props,a)){for(b=k.props[a],h(b)||(b=[b]),f=b.length,e=0;e<f;++e)c=b[e],j=c.indexOf("[VER]"),j>=0&&(c=c.substring(0,j)+"([\\w._\\+]+)"+c.substring(j+5)),b[e]=new RegExp(c,"i");k.props[a]=b}d(k.oss),d(k.phones),d(k.tablets),d(k.uas),d(k.utils),k.oss0={WindowsPhoneOS:k.oss.WindowsPhoneOS,WindowsMobileOS:k.oss.WindowsMobileOS}}(),g.findMatch=function(a,b){for(var c in a)if(i.call(a,c)&&a[c].test(b))return c;return null},g.findMatches=function(a,b){var c=[];for(var d in a)i.call(a,d)&&a[d].test(b)&&c.push(d);return c},g.getVersionStr=function(a,b){var c,d,e,f,h=g.mobileDetectRules.props;if(i.call(h,a))for(c=h[a],e=c.length,d=0;d<e;++d)if(f=c[d].exec(b),null!==f)return f[1];return null},g.getVersion=function(a,b){var c=g.getVersionStr(a,b);return c?g.prepareVersionNo(c):NaN},g.prepareVersionNo=function(a){var b;return b=a.split(/[a-z._ \/\-]/i),1===b.length&&(a=b[0]),b.length>1&&(a=b[0]+".",b.shift(),a+=b.join("")),Number(a)},g.isMobileFallback=function(a){return g.detectMobileBrowsers.fullPattern.test(a)||g.detectMobileBrowsers.shortPattern.test(a.substr(0,4))},g.isTabletFallback=function(a){return g.detectMobileBrowsers.tabletPattern.test(a)},g.prepareDetectionCache=function(a,c,d){if(a.mobile===b){var e,h,i;return(h=g.findMatch(g.mobileDetectRules.tablets,c))?(a.mobile=a.tablet=h,void(a.phone=null)):(e=g.findMatch(g.mobileDetectRules.phones,c))?(a.mobile=a.phone=e,void(a.tablet=null)):void(g.isMobileFallback(c)?(i=f.isPhoneSized(d),i===b?(a.mobile=g.FALLBACK_MOBILE,a.tablet=a.phone=null):i?(a.mobile=a.phone=g.FALLBACK_PHONE,a.tablet=null):(a.mobile=a.tablet=g.FALLBACK_TABLET,a.phone=null)):g.isTabletFallback(c)?(a.mobile=a.tablet=g.FALLBACK_TABLET,a.phone=null):a.mobile=a.tablet=a.phone=null)}},g.mobileGrade=function(a){var b=null!==a.mobile();return a.os("iOS")&&a.version("iPad")>=4.3||a.os("iOS")&&a.version("iPhone")>=3.1||a.os("iOS")&&a.version("iPod")>=3.1||a.version("Android")>2.1&&a.is("Webkit")||a.version("Windows Phone OS")>=7||a.is("BlackBerry")&&a.version("BlackBerry")>=6||a.match("Playbook.*Tablet")||a.version("webOS")>=1.4&&a.match("Palm|Pre|Pixi")||a.match("hp.*TouchPad")||a.is("Firefox")&&a.version("Firefox")>=12||a.is("Chrome")&&a.is("AndroidOS")&&a.version("Android")>=4||a.is("Skyfire")&&a.version("Skyfire")>=4.1&&a.is("AndroidOS")&&a.version("Android")>=2.3||a.is("Opera")&&a.version("Opera Mobi")>11&&a.is("AndroidOS")||a.is("MeeGoOS")||a.is("Tizen")||a.is("Dolfin")&&a.version("Bada")>=2||(a.is("UC Browser")||a.is("Dolfin"))&&a.version("Android")>=2.3||a.match("Kindle Fire")||a.is("Kindle")&&a.version("Kindle")>=3||a.is("AndroidOS")&&a.is("NookTablet")||a.version("Chrome")>=11&&!b||a.version("Safari")>=5&&!b||a.version("Firefox")>=4&&!b||a.version("MSIE")>=7&&!b||a.version("Opera")>=10&&!b?"A":a.os("iOS")&&a.version("iPad")<4.3||a.os("iOS")&&a.version("iPhone")<3.1||a.os("iOS")&&a.version("iPod")<3.1||a.is("Blackberry")&&a.version("BlackBerry")>=5&&a.version("BlackBerry")<6||a.version("Opera Mini")>=5&&a.version("Opera Mini")<=6.5&&(a.version("Android")>=2.3||a.is("iOS"))||a.match("NokiaN8|NokiaC7|N97.*Series60|Symbian/3")||a.version("Opera Mobi")>=11&&a.is("SymbianOS")?"B":(a.version("BlackBerry")<5||a.match("MSIEMobile|Windows CE.*Mobile")||a.version("Windows Mobile")<=5.2,"C")},g.detectOS=function(a){return g.findMatch(g.mobileDetectRules.oss0,a)||g.findMatch(g.mobileDetectRules.oss,a)},g.getDeviceSmallerSide=function(){return window.screen.width<window.screen.height?window.screen.width:window.screen.height},f.prototype={constructor:f,mobile:function(){return g.prepareDetectionCache(this._cache,this.ua,this.maxPhoneWidth),this._cache.mobile},phone:function(){return g.prepareDetectionCache(this._cache,this.ua,this.maxPhoneWidth),this._cache.phone},tablet:function(){return g.prepareDetectionCache(this._cache,this.ua,this.maxPhoneWidth),this._cache.tablet},userAgent:function(){return this._cache.userAgent===b&&(this._cache.userAgent=g.findMatch(g.mobileDetectRules.uas,this.ua)),this._cache.userAgent},userAgents:function(){return this._cache.userAgents===b&&(this._cache.userAgents=g.findMatches(g.mobileDetectRules.uas,this.ua)),this._cache.userAgents},os:function(){return this._cache.os===b&&(this._cache.os=g.detectOS(this.ua)),this._cache.os},version:function(a){return g.getVersion(a,this.ua)},versionStr:function(a){return g.getVersionStr(a,this.ua)},is:function(b){return c(this.userAgents(),b)||a(b,this.os())||a(b,this.phone())||a(b,this.tablet())||c(g.findMatches(g.mobileDetectRules.utils,this.ua),b)},match:function(a){return a instanceof RegExp||(a=new RegExp(a,"i")),a.test(this.ua)},isPhoneSized:function(a){return f.isPhoneSized(a||this.maxPhoneWidth)},mobileGrade:function(){return this._cache.grade===b&&(this._cache.grade=g.mobileGrade(this)),this._cache.grade}},"undefined"!=typeof window&&window.screen?f.isPhoneSized=function(a){return a<0?b:g.getDeviceSmallerSide()<=a}:f.isPhoneSized=function(){},f._impl=g,f.version="1.4.5 2021-03-13",f})}(function(a){if("undefined"!=typeof module&&module.exports)return function(a){module.exports=a()};if("function"==typeof define&&define.amd)return define;if("undefined"!=typeof window)return function(a){window.MobileDetect=a()};throw new Error("unknown environment")}());var ai_lists=!0,ai_block_class_def="code-block"; if("undefined"!=typeof ai_lists){function X(b,e){for(var p=[];b=b.previousElementSibling;)("undefined"==typeof e||b.matches(e))&&p.push(b);return p}function fa(b,e){for(var p=[];b=b.nextElementSibling;)("undefined"==typeof e||b.matches(e))&&p.push(b);return p}var host_regexp=RegExp(":\\/\\/(.[^/:]+)","i");function ha(b){b=b.match(host_regexp);return null!=b&&1<b.length&&"string"===typeof b[1]&&0<b[1].length?b[1].toLowerCase():null}function Q(b){return b.includes(":")?(b=b.split(":"),1E3*(3600*parseInt(b[0])+ 60*parseInt(b[1])+parseInt(b[2]))):null}function Y(b){try{var e=Date.parse(b);isNaN(e)&&(e=null)}catch(p){e=null}if(null==e&&b.includes(" ")){b=b.split(" ");try{e=Date.parse(b[0]),e+=Q(b[1]),isNaN(e)&&(e=null)}catch(p){e=null}}return e}function Z(){null==document.querySelector("#ai-iab-tcf-bar")&&null==document.querySelector(".ai-list-manual")||"function"!=typeof __tcfapi||"function"!=typeof ai_load_blocks||"undefined"!=typeof ai_iab_tcf_callback_installed||(__tcfapi("addEventListener",2,function(b, e){e&&"useractioncomplete"===b.eventStatus&&(ai_tcData=b,ai_load_blocks(),b=document.querySelector("#ai-iab-tcf-status"),null!=b&&(b.textContent="IAB TCF 2.0 DATA LOADED"),b=document.querySelector("#ai-iab-tcf-bar"),null!=b&&(b.classList.remove("status-error"),b.classList.add("status-ok")))}),ai_iab_tcf_callback_installed=!0)}ai_process_lists=function(b){function e(a,d,k){if(0==a.length){if("!@!"==k)return!0;d!=k&&("true"==k.toLowerCase()?k=!0:"false"==k.toLowerCase()&&(k=!1));return d==k}if("object"!= typeof d&&"array"!=typeof d)return!1;var l=a[0];a=a.slice(1);if("*"==l)for(let [,n]of Object.entries(d)){if(e(a,n,k))return!0}else if(l in d)return e(a,d[l],k);return!1}function p(a,d,k){if("object"!=typeof a||-1==d.indexOf("["))return!1;d=d.replace(/]| /gi,"").split("[");return e(d,a,k)}function z(){if("function"==typeof __tcfapi){var a=document.querySelector("#ai-iab-tcf-status"),d=document.querySelector("#ai-iab-tcf-bar");null!=a&&(a.textContent="IAB TCF 2.0 DETECTED");__tcfapi("getTCData",2,function(k, l){l?(null!=d&&d.classList.add("status-ok"),"tcloaded"==k.eventStatus||"useractioncomplete"==k.eventStatus)?(ai_tcData=k,k.gdprApplies?null!=a&&(a.textContent="IAB TCF 2.0 DATA LOADED"):null!=a&&(a.textContent="IAB TCF 2.0 GDPR DOES NOT APPLY"),null!=d&&(d.classList.remove("status-error"),d.classList.add("status-ok")),setTimeout(function(){ai_process_lists()},10)):"cmpuishown"==k.eventStatus&&(ai_cmpuishown=!0,null!=a&&(a.textContent="IAB TCF 2.0 CMP UI SHOWN"),null!=d&&(d.classList.remove("status-error"), d.classList.add("status-ok"))):(null!=a&&(a.textContent="IAB TCF 2.0 __tcfapi getTCData failed"),null!=d&&(d.classList.remove("status-ok"),d.classList.add("status-error")))})}}function C(a){"function"==typeof __tcfapi?(ai_tcfapi_found=!0,"undefined"==typeof ai_iab_tcf_callback_installed&&Z(),"undefined"==typeof ai_tcData_requested&&(ai_tcData_requested=!0,z(),cookies_need_tcData=!0)):a&&("undefined"==typeof ai_tcfapi_found&&(ai_tcfapi_found=!1,setTimeout(function(){ai_process_lists()},10)),a=document.querySelector("#ai-iab-tcf-status"), null!=a&&(a.textContent="IAB TCF 2.0 MISSING: __tcfapi function not found"),a=document.querySelector("#ai-iab-tcf-bar"),null!=a&&(a.classList.remove("status-ok"),a.classList.add("status-error")))}if(null==b)b=document.querySelectorAll("div.ai-list-data, meta.ai-list-data");else{window.jQuery&&window.jQuery.fn&&b instanceof jQuery&&(b=Array.prototype.slice.call(b));var x=[];b.forEach((a,d)=>{a.matches(".ai-list-data")?x.push(a):(a=a.querySelectorAll(".ai-list-data"),a.length&&a.forEach((k,l)=>{x.push(k)}))}); b=x}if(b.length){b.forEach((a,d)=>{a.classList.remove("ai-list-data")});var L=ia(window.location.search);if(null!=L.referrer)var A=L.referrer;else A=document.referrer,""!=A&&(A=ha(A));var R=window.navigator.userAgent,S=R.toLowerCase(),aa=navigator.language,M=aa.toLowerCase();if("undefined"!==typeof MobileDetect)var ba=new MobileDetect(R);b.forEach((a,d)=>{var k=document.cookie.split(";");k.forEach(function(f,h){k[h]=f.trim()});d=a.closest("div."+ai_block_class_def);var l=!0;if(a.hasAttribute("referer-list")){var n= a.getAttribute("referer-list");n=b64d(n).split(",");var v=a.getAttribute("referer-list-type"),E=!1;n.every((f,h)=>{f=f.trim();if(""==f)return!0;if("*"==f.charAt(0))if("*"==f.charAt(f.length-1)){if(f=f.substr(1,f.length-2),-1!=A.indexOf(f))return E=!0,!1}else{if(f=f.substr(1),A.substr(-f.length)==f)return E=!0,!1}else if("*"==f.charAt(f.length-1)){if(f=f.substr(0,f.length-1),0==A.indexOf(f))return E=!0,!1}else if("#"==f){if(""==A)return E=!0,!1}else if(f==A)return E=!0,!1;return!0});var r=E;switch(v){case "B":r&& (l=!1);break;case "W":r||(l=!1)}}if(l&&a.hasAttribute("client-list")&&"undefined"!==typeof ba)switch(n=a.getAttribute("client-list"),n=b64d(n).split(","),v=a.getAttribute("client-list-type"),r=!1,n.every((f,h)=>{if(""==f.trim())return!0;f.split("&&").every((c,t)=>{t=!0;var w=!1;for(c=c.trim();"!!"==c.substring(0,2);)t=!t,c=c.substring(2);"language:"==c.substring(0,9)&&(w=!0,c=c.substring(9).toLowerCase());var q=!1;w?"*"==c.charAt(0)?"*"==c.charAt(c.length-1)?(c=c.substr(1,c.length-2).toLowerCase(), -1!=M.indexOf(c)&&(q=!0)):(c=c.substr(1).toLowerCase(),M.substr(-c.length)==c&&(q=!0)):"*"==c.charAt(c.length-1)?(c=c.substr(0,c.length-1).toLowerCase(),0==M.indexOf(c)&&(q=!0)):c==M&&(q=!0):"*"==c.charAt(0)?"*"==c.charAt(c.length-1)?(c=c.substr(1,c.length-2).toLowerCase(),-1!=S.indexOf(c)&&(q=!0)):(c=c.substr(1).toLowerCase(),S.substr(-c.length)==c&&(q=!0)):"*"==c.charAt(c.length-1)?(c=c.substr(0,c.length-1).toLowerCase(),0==S.indexOf(c)&&(q=!0)):ba.is(c)&&(q=!0);return(r=q?t:!t)?!0:!1});return r? !1:!0}),v){case "B":r&&(l=!1);break;case "W":r||(l=!1)}var N=n=!1;for(v=1;2>=v;v++)if(l){switch(v){case 1:var g=a.getAttribute("cookie-list");break;case 2:g=a.getAttribute("parameter-list")}if(null!=g){g=b64d(g);switch(v){case 1:var y=a.getAttribute("cookie-list-type");break;case 2:y=a.getAttribute("parameter-list-type")}g=g.replace("tcf-gdpr","tcf-v2[gdprApplies]=true");g=g.replace("tcf-no-gdpr","tcf-v2[gdprApplies]=false");g=g.replace("tcf-google","tcf-v2[vendor][consents][755]=true && tcf-v2[purpose][consents][1]=true"); g=g.replace("tcf-no-google","!!tcf-v2[vendor][consents][755]");g=g.replace("tcf-media.net","tcf-v2[vendor][consents][142]=true && tcf-v2[purpose][consents][1]=true");g=g.replace("tcf-no-media.net","!!tcf-v2[vendor][consents][142]");g=g.replace("tcf-amazon","tcf-v2[vendor][consents][793]=true && tcf-v2[purpose][consents][1]=true");g=g.replace("tcf-no-amazon","!!tcf-v2[vendor][consents][793]");g=g.replace("tcf-ezoic","tcf-v2[vendor][consents][347]=true && tcf-v2[purpose][consents][1]=true");g=g.replace("tcf-no-ezoic", "!!tcf-v2[vendor][consents][347]");var F=g.split(","),ca=[];k.forEach(function(f){f=f.split("=");try{var h=JSON.parse(decodeURIComponent(f[1]))}catch(c){h=decodeURIComponent(f[1])}ca[f[0]]=h});r=!1;var I=a;F.every((f,h)=>{f.split("&&").every((c,t)=>{t=!0;for(c=c.trim();"!!"==c.substring(0,2);)t=!t,c=c.substring(2);var w=c,q="!@!",T="tcf-v2"==w&&"!@!"==q,B=-1!=c.indexOf("["),J=0==c.indexOf("tcf-v2")||0==c.indexOf("euconsent-v2");J=J&&(B||T);-1!=c.indexOf("=")&&(q=c.split("="),w=q[0],q=q[1],B=-1!=w.indexOf("["), J=(J=0==w.indexOf("tcf-v2")||0==w.indexOf("euconsent-v2"))&&(B||T));if(J)document.querySelector("#ai-iab-tcf-status"),B=document.querySelector("#ai-iab-tcf-bar"),null!=B&&(B.style.display="block"),T&&"boolean"==typeof ai_tcfapi_found?r=ai_tcfapi_found?t:!t:"object"==typeof ai_tcData?(null!=B&&B.classList.add("status-ok"),w=w.replace(/]| /gi,"").split("["),w.shift(),r=(w=e(w,ai_tcData,q))?t:!t):"undefined"==typeof ai_tcfapi_found&&(I.classList.add("ai-list-data"),N=!0,"function"==typeof __tcfapi?C(!1): "undefined"==typeof ai_tcData_retrying&&(ai_tcData_retrying=!0,setTimeout(function(){"function"==typeof __tcfapi?C(!1):setTimeout(function(){"function"==typeof __tcfapi?C(!1):setTimeout(function(){C(!0)},3E3)},1E3)},600)));else if(B)r=(w=p(ca,w,q))?t:!t;else{var U=!1;"!@!"==q?k.every(function(ja){return ja.split("=")[0]==c?(U=!0,!1):!0}):U=-1!=k.indexOf(c);r=U?t:!t}return r?!0:!1});return r?!1:!0});r&&(N=!1,I.classList.remove("ai-list-data"));switch(y){case "B":r&&(l=!1);break;case "W":r||(l=!1)}}}a.classList.contains("ai-list-manual")&& (l?(I.classList.remove("ai-list-data"),I.classList.remove("ai-list-manual")):(n=!0,I.classList.add("ai-list-data")));(l||!n&&!N)&&a.hasAttribute("data-debug-info")&&(g=document.querySelector("."+a.dataset.debugInfo),null!=g&&(g=g.parentElement,null!=g&&g.classList.contains("ai-debug-info")&&g.remove()));y=X(a,".ai-debug-bar.ai-debug-lists");var ka=""==A?"#":A;0!=y.length&&y.forEach((f,h)=>{h=f.querySelector(".ai-debug-name.ai-list-info");null!=h&&(h.textContent=ka,h.title=R+"\n"+aa);h=f.querySelector(".ai-debug-name.ai-list-status"); null!=h&&(h.textContent=l?ai_front.visible:ai_front.hidden)});g=!1;if(l&&a.hasAttribute("scheduling-start")&&a.hasAttribute("scheduling-end")&&a.hasAttribute("scheduling-days")){var u=a.getAttribute("scheduling-start");v=a.getAttribute("scheduling-end");y=a.getAttribute("scheduling-days");g=!0;u=b64d(u);F=b64d(v);var V=parseInt(a.getAttribute("scheduling-fallback")),O=parseInt(a.getAttribute("gmt"));if(u.includes("-")||F.includes("-"))P=Y(u)+O,K=Y(F)+O;else var P=Q(u),K=Q(F);P??=0;K??=0;var W=b64d(y).split(","); y=a.getAttribute("scheduling-type");var D=(new Date).getTime()+O;v=new Date(D);var G=v.getDay();0==G?G=6:G--;u.includes("-")||F.includes("-")||(u=(new Date(v.getFullYear(),v.getMonth(),v.getDate())).getTime()+O,D-=u,0>D&&(D+=864E5));scheduling_start_date_ok=D>=P;scheduling_end_date_ok=0==K||D<K;u=scheduling_start_date_ok&&scheduling_end_date_ok&&W.includes(G.toString());switch(y){case "B":u=!u}u||(l=!1);var la=v.toISOString().split(".")[0].replace("T"," ");y=X(a,".ai-debug-bar.ai-debug-scheduling"); 0!=y.length&&y.forEach((f,h)=>{h=f.querySelector(".ai-debug-name.ai-scheduling-info");null!=h&&(h.textContent=la+" "+G+" current_time: "+Math.floor(D.toString()/1E3)+" start_date:"+Math.floor(P/1E3).toString()+"=>"+scheduling_start_date_ok.toString()+" end_date:"+Math.floor(K/1E3).toString()+"=>"+scheduling_end_date_ok.toString()+" days:"+W.toString()+"=>"+W.includes(G.toString()).toString());h=f.querySelector(".ai-debug-name.ai-scheduling-status");null!=h&&(h.textContent=l?ai_front.visible:ai_front.hidden); l||0==V||(f.classList.remove("ai-debug-scheduling"),f.classList.add("ai-debug-fallback"),h=f.querySelector(".ai-debug-name.ai-scheduling-status"),null!=h&&(h.textContent=ai_front.fallback+" = "+V))})}if(n||!l&&N)return!0;a.style.visibility="";a.style.position="";a.style.width="";a.style.height="";a.style.zIndex="";if(l){if(null!=d&&(d.style.visibility="",d.classList.contains("ai-remove-position")&&(d.style.position="")),a.hasAttribute("data-code")){n=b64d(a.dataset.code);u=document.createRange(); g=!0;try{H=u.createContextualFragment(n)}catch(f){g=!1}g&&(null!=a.closest("head")?(a.parentNode.insertBefore(H,a.nextSibling),a.remove()):a.append(H));da(a)}}else if(g&&!u&&0!=V){null!=d&&(d.style.visibility="",d.classList.contains("ai-remove-position")&&d.css({position:""}));n=fa(a,".ai-fallback");0!=n.length&&n.forEach((f,h)=>{f.classList.remove("ai-fallback")});if(a.hasAttribute("data-fallback-code")){n=b64d(a.dataset.fallbackCode);u=document.createRange();g=!0;try{var H=u.createContextualFragment(n)}catch(f){g= !1}g&&a.append(H);da(a)}else a.style.display="none",null!=d&&null==d.querySelector(".ai-debug-block")&&d.hasAttribute("style")&&-1==d.getAttribute("style").indexOf("height:")&&(d.style.display="none");null!=d&&d.hasAttribute("data-ai")&&(d.getAttribute("data-ai"),a.hasAttribute("fallback-tracking")&&(H=a.getAttribute("fallback-tracking"),d.setAttribute("data-ai-"+a.getAttribute("fallback_level"),H)))}else a.style.display="none",null!=d&&(d.removeAttribute("data-ai"),d.classList.remove("ai-track"), null!=d.querySelector(".ai-debug-block")?(d.style.visibility="",d.classList.remove("ai-close"),d.classList.contains("ai-remove-position")&&(d.style.position="")):d.hasAttribute("style")&&-1==d.getAttribute("style").indexOf("height:")&&(d.style.display="none"));a.setAttribute("data-code","");a.setAttribute("data-fallback-code","");null!=d&&d.classList.remove("ai-list-block")})}};function ea(b){b=`; ${document.cookie}`.split(`; ${b}=`);if(2===b.length)return b.pop().split(";").shift()}function ma(b, e,p){ea(b)&&(document.cookie=b+"="+(e?";path="+e:"")+(p?";domain="+p:"")+";expires=Thu, 01 Jan 1970 00:00:01 GMT")}function m(b){ea(b)&&(ma(b,"/",window.location.hostname),document.cookie=b+"=; Path=/; Expires=Thu, 01 Jan 1970 00:00:01 GMT;")}(function(b){"complete"===document.readyState||"loading"!==document.readyState&&!document.documentElement.doScroll?b():document.addEventListener("DOMContentLoaded",b)})(function(){setTimeout(function(){ai_process_lists();setTimeout(function(){Z();if("function"== typeof ai_load_blocks){document.addEventListener("cmplzEnableScripts",e);document.addEventListener("cmplz_event_marketing",e);function e(p){"cmplzEnableScripts"!=p.type&&"all"!==p.consentLevel||ai_load_blocks()}}},50);var b=document.querySelector(".ai-debug-page-type");null!=b&&b.addEventListener("dblclick",e=>{e=document.querySelector("#ai-iab-tcf-status");null!=e&&(e.textContent="CONSENT COOKIES");e=document.querySelector("#ai-iab-tcf-bar");null!=e&&(e.style.display="block")});b=document.querySelector("#ai-iab-tcf-bar"); null!=b&&b.addEventListener("click",e=>{m("euconsent-v2");m("__lxG__consent__v2");m("__lxG__consent__v2_daisybit");m("__lxG__consent__v2_gdaisybit");m("CookieLawInfoConsent");m("cookielawinfo-checkbox-advertisement");m("cookielawinfo-checkbox-analytics");m("cookielawinfo-checkbox-necessary");m("complianz_policy_id");m("complianz_consent_status");m("cmplz_marketing");m("cmplz_consent_status");m("cmplz_preferences");m("cmplz_statistics-anonymous");m("cmplz_choice");m("cmplz_banner-status");m("cmplz_functional"); m("cmplz_policy_id");m("cmplz_statistics");m("moove_gdpr_popup");m("real_cookie_banner-blog:1-tcf");m("real_cookie_banner-blog:1");e=document.querySelector("#ai-iab-tcf-status");null!=e&&(e.textContent="CONSENT COOKIES DELETED")})},5)});function da(b){setTimeout(function(){"function"==typeof ai_process_rotations_in_element&&ai_process_rotations_in_element(b);"function"==typeof ai_process_lists&&ai_process_lists();"function"==typeof ai_process_ip_addresses&&ai_process_ip_addresses();"function"==typeof ai_process_filter_hooks&& ai_process_filter_hooks();"function"==typeof ai_adb_process_blocks&&ai_adb_process_blocks(b);"function"==typeof ai_process_impressions&&1==ai_tracking_finished&&ai_process_impressions();"function"==typeof ai_install_click_trackers&&1==ai_tracking_finished&&ai_install_click_trackers();"function"==typeof ai_install_close_buttons&&ai_install_close_buttons(document)},5)}function ia(b){var e=b?b.split("?")[1]:window.location.search.slice(1);b={};if(e){e=e.split("#")[0];e=e.split("&");for(var p=0;p<e.length;p++){var z= e[p].split("="),C=void 0,x=z[0].replace(/\[\d*\]/,function(L){C=L.slice(1,-1);return""});z="undefined"===typeof z[1]?"":z[1];x=x.toLowerCase();z=z.toLowerCase();b[x]?("string"===typeof b[x]&&(b[x]=[b[x]]),"undefined"===typeof C?b[x].push(z):b[x][C]=z):b[x]=z}}return b}}; ai_run_652362550520 = function(){ ai_document_write=document.write;document.write=function(a){"interactive"==document.readyState?(console.error("document.write called after page load: ",a),"undefined"!=typeof ai_js_errors&&ai_js_errors.push(["document.write called after page load",a,0])):ai_document_write.call(document,a)}; ai_insert ('before', '#nv-footer', b64d ('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')); }; if (document.readyState === 'complete' || (document.readyState !== 'loading' && !document.documentElement.doScroll)) ai_run_652362550520 (); else document.addEventListener ('DOMContentLoaded', ai_run_652362550520); ai_js_code = true; </script> <div class="sticky-share-buttons"> <ul class="entry-meta-social-links-list"> <li><a data-wpel-link="external" href="https://www.linkedin.com/sharing/share-offsite/?url=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-linkedin" target="_blank" rel="follow external noopener">L</a></li> <li><a data-wpel-link="external" href="https://twitter.com/intent/tweet?text=Customizing+AI+Models%3A+Train+Character+Detection+and+Recognition+Models+with+NVIDIA+TAO+%7C+NVIDIA+Technical+Blog+https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-twitter" target="_blank" rel="follow external noopener">T</a></li> <li><a data-wpel-link="external" href="https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-facebook" target="_blank" rel="follow external noopener">F</a></li> <li><a data-wpel-link="external" href="https://www.reddit.com/submit?url=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F&title=Customizing+AI+Models%3A+Train+Character+Detection+and+Recognition+Models+with+NVIDIA+TAO+%7C+NVIDIA+Technical+Blog" class="for-reddit" target="_blank" rel="follow external noopener">R</a></li> <li><a href="mailto:?subject=I'd like to share a link with you&body=https%3A%2F%2Fdeveloper.nvidia.com%2Fblog%2Fcreate-custom-character-detection-and-recognition-models-with-nvidia-tao-part-1%2F" class="for-mail">E</a></li> </ul> </div> <script type="text/javascript">window.NREUM||(NREUM={});NREUM.info={"beacon":"bam.nr-data.net","licenseKey":"NRJS-2b0defbeb07862260b2","applicationID":"745178505","transactionName":"Z10BbBRRW0QCAExfWF4XIlsSWVpZTApWUlJI","queueTime":0,"applicationTime":1259,"atts":"SxoCGlxLSEo=","errorBeacon":"bam.nr-data.net","agent":""}</script></body> </html>