Buckets:

hf-doc-build
/

doc-dev

hf-doc-build/doc-dev / course /pr_1159 /fr /_app /immutable /nodes /29.129bb129.js

rtrm's picture

3 months ago

17.2 kB

	import{s as Ue,o as $e}from"../chunks/scheduler.893fe8c9.js";import{S as we,i as Ce,e as M,s as c,c as g,h as _e,a as y,d as a,b as o,f as Te,g as J,j as C,k as X,l as Ze,m as r,n as T,o as f,q as ke,t as h,p as k,r as je}from"../chunks/index.2d09ebb4.js";import{C as ve,H as We,E as Ve}from"../chunks/MermaidChart.svelte_svelte_type_style_lang.1d939c26.js";import{C as L}from"../chunks/CodeBlock.53a6f786.js";import{C as xe}from"../chunks/CourseFloatingBanner.2900b001.js";import{F as Ne}from"../chunks/FrameworkSwitchCourse.f49b9dc4.js";function ze(j){let n,i;return n=new xe({props:{chapter:4,classNames:"absolute z-10 right-0 top-0",notebooks:[{label:"English",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/master/course/en/chapter4/section2_tf.ipynb"},{label:"Français",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/master/course/fr/chapter4/section2_tf.ipynb"},{label:"English",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/master/course/en/chapter4/section2_tf.ipynb"},{label:"Français",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/master/course/fr/chapter4/section2_tf.ipynb"}]}}),{c(){g(n.$$.fragment)},l(t){J(n.$$.fragment,t)},m(t,d){T(n,t,d),i=!0},i(t){i\|\|(h(n.$$.fragment,t),i=!0)},o(t){f(n.$$.fragment,t),i=!1},d(t){k(n,t)}}}function Ee(j){let n,i;return n=new xe({props:{chapter:4,classNames:"absolute z-10 right-0 top-0",notebooks:[{label:"English",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/master/course/en/chapter4/section2_pt.ipynb"},{label:"Français",value:"https://colab.research.google.com/github/huggingface/notebooks/blob/master/course/fr/chapter4/section2_pt.ipynb"},{label:"English",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/master/course/en/chapter4/section2_pt.ipynb"},{label:"Français",value:"https://studiolab.sagemaker.aws/import/github/huggingface/notebooks/blob/master/course/fr/chapter4/section2_pt.ipynb"}]}}),{c(){g(n.$$.fragment)},l(t){J(n.$$.fragment,t)},m(t,d){T(n,t,d),i=!0},i(t){i\|\|(h(n.$$.fragment,t),i=!0)},o(t){f(n.$$.fragment,t),i=!1},d(t){k(n,t)}}}function Fe(j){let n,i,t,d='Cependant, nous recommandons d’utiliser les classes <a href="https://huggingface.co/transformers/model_doc/auto.html?highlight=auto#auto-classes" rel="nofollow"><code>TFAuto</code></a> à la place, car elles sont par conception indépendantes de l’architecture. Alors que l’exemple de code précédent limite les utilisateurs aux <em>checkpoints</em> chargeables dans l’architecture CamemBERT, l’utilisation des classes <code>TFAuto</code> facilite le changement de <em>checkpoint</em> :',u,m,b;return n=new L({props:{code:"ZnJvbSUyMHRyYW5zZm9ybWVycyUyMGltcG9ydCUyMENhbWVtYmVydFRva2VuaXplciUyQyUyMFRGQ2FtZW1iZXJ0Rm9yTWFza2VkTE0lMEElMEF0b2tlbml6ZXIlMjAlM0QlMjBDYW1lbWJlcnRUb2tlbml6ZXIuZnJvbV9wcmV0cmFpbmVkKCUyMmNhbWVtYmVydC1iYXNlJTIyKSUwQW1vZGVsJTIwJTNEJTIwVEZDYW1lbWJlcnRGb3JNYXNrZWRMTS5mcm9tX3ByZXRyYWluZWQoJTIyY2FtZW1iZXJ0LWJhc2UlMjIp",highlighted:`<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> CamembertTokenizer, TFCamembertForMaskedLM

	tokenizer = CamembertTokenizer.from_pretrained(<span class="hljs-string">"camembert-base"</span>)
	model = TFCamembertForMaskedLM.from_pretrained(<span class="hljs-string">"camembert-base"</span>)`,wrap:!1}}),m=new L({props:{code:"ZnJvbSUyMHRyYW5zZm9ybWVycyUyMGltcG9ydCUyMEF1dG9Ub2tlbml6ZXIlMkMlMjBURkF1dG9Nb2RlbEZvck1hc2tlZExNJTBBJTBBdG9rZW5pemVyJTIwJTNEJTIwQXV0b1Rva2VuaXplci5mcm9tX3ByZXRyYWluZWQoJTIyY2FtZW1iZXJ0LWJhc2UlMjIpJTBBbW9kZWwlMjAlM0QlMjBURkF1dG9Nb2RlbEZvck1hc2tlZExNLmZyb21fcHJldHJhaW5lZCglMjJjYW1lbWJlcnQtYmFzZSUyMik=",highlighted:`<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoTokenizer, TFAutoModelForMaskedLM

	tokenizer = AutoTokenizer.from_pretrained(<span class="hljs-string">"camembert-base"</span>)
	model = TFAutoModelForMaskedLM.from_pretrained(<span class="hljs-string">"camembert-base"</span>)`,wrap:!1}}),{c(){g(n.$$.fragment),i=c(),t=M("p"),t.innerHTML=d,u=c(),g(m.$$.fragment)},l(s){J(n.$$.fragment,s),i=o(s),t=y(s,"P",{"data-svelte-h":!0}),C(t)!=="svelte-83gnjn"&&(t.innerHTML=d),u=o(s),J(m.$$.fragment,s)},m(s,p){T(n,s,p),r(s,i,p),r(s,t,p),r(s,u,p),T(m,s,p),b=!0},i(s){b\|\|(h(n.$$.fragment,s),h(m.$$.fragment,s),b=!0)},o(s){f(n.$$.fragment,s),f(m.$$.fragment,s),b=!1},d(s){s&&(a(i),a(t),a(u)),k(n,s),k(m,s)}}}function Qe(j){let n,i,t,d='Cependant, nous recommandons d’utiliser les classes <a href="https://huggingface.co/transformers/model_doc/auto.html?highlight=auto#auto-classes" rel="nofollow"><code>Auto</code></a> à la place, car elles sont par conception indépendantes de l’architecture. Alors que l’exemple de code précédent limite les utilisateurs aux <em>checkpoints</em> chargeables dans l’architecture CamemBERT, l’utilisation des classes <code>Auto</code> facilite le changement de <em>checkpoint</em> :',u,m,b;return n=new L({props:{code:"ZnJvbSUyMHRyYW5zZm9ybWVycyUyMGltcG9ydCUyMENhbWVtYmVydFRva2VuaXplciUyQyUyMENhbWVtYmVydEZvck1hc2tlZExNJTBBJTBBdG9rZW5pemVyJTIwJTNEJTIwQ2FtZW1iZXJ0VG9rZW5pemVyLmZyb21fcHJldHJhaW5lZCglMjJjYW1lbWJlcnQtYmFzZSUyMiklMEFtb2RlbCUyMCUzRCUyMENhbWVtYmVydEZvck1hc2tlZExNLmZyb21fcHJldHJhaW5lZCglMjJjYW1lbWJlcnQtYmFzZSUyMik=",highlighted:`<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> CamembertTokenizer, CamembertForMaskedLM

	tokenizer = CamembertTokenizer.from_pretrained(<span class="hljs-string">"camembert-base"</span>)
	model = CamembertForMaskedLM.from_pretrained(<span class="hljs-string">"camembert-base"</span>)`,wrap:!1}}),m=new L({props:{code:"ZnJvbSUyMHRyYW5zZm9ybWVycyUyMGltcG9ydCUyMEF1dG9Ub2tlbml6ZXIlMkMlMjBBdXRvTW9kZWxGb3JNYXNrZWRMTSUwQSUwQXRva2VuaXplciUyMCUzRCUyMEF1dG9Ub2tlbml6ZXIuZnJvbV9wcmV0cmFpbmVkKCUyMmNhbWVtYmVydC1iYXNlJTIyKSUwQW1vZGVsJTIwJTNEJTIwQXV0b01vZGVsRm9yTWFza2VkTE0uZnJvbV9wcmV0cmFpbmVkKCUyMmNhbWVtYmVydC1iYXNlJTIyKQ==",highlighted:`<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoTokenizer, AutoModelForMaskedLM

	tokenizer = AutoTokenizer.from_pretrained(<span class="hljs-string">"camembert-base"</span>)
	model = AutoModelForMaskedLM.from_pretrained(<span class="hljs-string">"camembert-base"</span>)`,wrap:!1}}),{c(){g(n.$$.fragment),i=c(),t=M("p"),t.innerHTML=d,u=c(),g(m.$$.fragment)},l(s){J(n.$$.fragment,s),i=o(s),t=y(s,"P",{"data-svelte-h":!0}),C(t)!=="svelte-8qvpr"&&(t.innerHTML=d),u=o(s),J(m.$$.fragment,s)},m(s,p){T(n,s,p),r(s,i,p),r(s,t,p),r(s,u,p),T(m,s,p),b=!0},i(s){b\|\|(h(n.$$.fragment,s),h(m.$$.fragment,s),b=!0)},o(s){f(n.$$.fragment,s),f(m.$$.fragment,s),b=!1},d(s){s&&(a(i),a(t),a(u)),k(n,s),k(m,s)}}}function Be(j){let n,i,t,d,u,m,b,s,p,S,x,U,q,W,ce="Le <em>Hub</em> rend simple la sélection d’un modèle et permet alors que celui-ci puisse être utilisé dans toute bibliothèque en aval en seulement quelques lignes de code. Voyons comment utiliser concrètement l’un de ces modèles et comment contribuer au développement de la communauté.",R,V,oe="Supposons que nous recherchions un modèle basé sur le français, capable de remplir des masques.",H,_,me='<img src="https://huggingface.co/datasets/huggingface-course/documentation-images/resolve/main/en/chapter4/camembert.gif" alt="Selecting the Camembert model." width="80%"/>',D,N,pe="Nous choisissons le <em>checkpoint</em> <code>camembert-base</code> pour essayer. L’identifiant <code>camembert-base</code> est tout ce dont nous avons besoin pour commencer à utiliser le modèle ! Comme vous l’avez vu dans les chapitres précédents, nous pouvons l’instancier en utilisant la fonction <code>pipeline()</code> :",P,z,K,E,O,F,ue="Comme vous pouvez le constater, le chargement d’un modèle dans un pipeline est extrêmement simple. La seule chose à laquelle vous devez faire attention est que le <em>checkpoint</em> choisi soit adapté à la tâche pour laquelle il va être utilisé. Par exemple, ici nous chargeons le <em>checkpoint</em> <code>camembert-base</code> dans le pipeline <code>fill-mask</code>, ce qui est tout à fait correct. Mais si nous chargerions ce <em>checkpoint</em> dans le pipeline <code>text-classification</code>, les résultats n’auraient aucun sens car la tête de <code>camembert-base</code> n’est pas adaptée à cette tâche ! Nous recommandons d’utiliser le sélecteur de tâche dans l’interface du <em>Hub</em> afin de sélectionner les <em>checkpoints</em> appropriés :",ee,Z,be='<img src="https://huggingface.co/datasets/huggingface-course/documentation-images/resolve/main/en/chapter4/tasks.png" alt="The task selector on the web interface." width="80%"/>',se,Q,de="Vous pouvez également instancier le <em>checkpoint</em> en utilisant directement l’architecture du modèle :",te,$,w,G,v,fe="<p>Lorsque vous utilisez un modèle pré-entraîné, assurez-vous de vérifier comment il a été entraîné, sur quels jeux de données, ses limites et ses biais. Toutes ces informations doivent être indiquées dans sa carte.</p>",le,B,ne,A,ae;u=new Ne({props:{fw:j[0]}}),b=new ve({props:{containerStyle:"float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"}}),p=new We({props:{title:"Utilisation de modèles pré-entraînés",local:"utilisation-de-modèles-pré-entraînés",headingTag:"h1"}});const he=[Ee,ze],I=[];function Me(e,l){return e[0]==="pt"?0:1}x=Me(j),U=I[x]=he[x](j),z=new L({props:{code:"ZnJvbSUyMHRyYW5zZm9ybWVycyUyMGltcG9ydCUyMHBpcGVsaW5lJTBBJTBBY2FtZW1iZXJ0X2ZpbGxfbWFzayUyMCUzRCUyMHBpcGVsaW5lKCUyMmZpbGwtbWFzayUyMiUyQyUyMG1vZGVsJTNEJTIyY2FtZW1iZXJ0LWJhc2UlMjIpJTBBcmVzdWx0cyUyMCUzRCUyMGNhbWVtYmVydF9maWxsX21hc2soJTIyTGUlMjBjYW1lbWJlcnQlMjBlc3QlMjAlM0NtYXNrJTNFJTIwJTNBKSUyMik=",highlighted:`<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> pipeline

	camembert_fill_mask = pipeline(<span class="hljs-string">"fill-mask"</span>, model=<span class="hljs-string">"camembert-base"</span>)
	results = camembert_fill_mask(<span class="hljs-string">"Le camembert est <mask> :)"</span>)`,wrap:!1}}),E=new L({props:{code:"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",highlighted:`[
	{<span class="hljs-string">'sequence'</span>: <span class="hljs-string">'Le camembert est délicieux :)'</span>, <span class="hljs-string">'score'</span>: <span class="hljs-number">0.49091005325317383</span>, <span class="hljs-string">'token'</span>: <span class="hljs-number">7200</span>, <span class="hljs-string">'token_str'</span>: <span class="hljs-string">'délicieux'</span>},
	{<span class="hljs-string">'sequence'</span>: <span class="hljs-string">'Le camembert est excellent :)'</span>, <span class="hljs-string">'score'</span>: <span class="hljs-number">0.1055697426199913</span>, <span class="hljs-string">'token'</span>: <span class="hljs-number">2183</span>, <span class="hljs-string">'token_str'</span>: <span class="hljs-string">'excellent'</span>},
	{<span class="hljs-string">'sequence'</span>: <span class="hljs-string">'Le camembert est succulent :)'</span>, <span class="hljs-string">'score'</span>: <span class="hljs-number">0.03453313186764717</span>, <span class="hljs-string">'token'</span>: <span class="hljs-number">26202</span>, <span class="hljs-string">'token_str'</span>: <span class="hljs-string">'succulent'</span>},
	{<span class="hljs-string">'sequence'</span>: <span class="hljs-string">'Le camembert est meilleur :)'</span>, <span class="hljs-string">'score'</span>: <span class="hljs-number">0.0330314114689827</span>, <span class="hljs-string">'token'</span>: <span class="hljs-number">528</span>, <span class="hljs-string">'token_str'</span>: <span class="hljs-string">'meilleur'</span>},
	{<span class="hljs-string">'sequence'</span>: <span class="hljs-string">'Le camembert est parfait :)'</span>, <span class="hljs-string">'score'</span>: <span class="hljs-number">0.03007650189101696</span>, <span class="hljs-string">'token'</span>: <span class="hljs-number">1654</span>, <span class="hljs-string">'token_str'</span>: <span class="hljs-string">'parfait'</span>}
	]`,wrap:!1}});const ye=[Qe,Fe],Y=[];function ge(e,l){return e[0]==="pt"?0:1}return $=ge(j),w=Y[$]=ye[$](j),B=new Ve({props:{source:"https://github.com/huggingface/course/blob/main/chapters/fr/chapter4/2.mdx"}}),{c(){n=M("meta"),i=c(),t=M("p"),d=c(),g(u.$$.fragment),m=c(),g(b.$$.fragment),s=c(),g(p.$$.fragment),S=c(),U.c(),q=c(),W=M("p"),W.innerHTML=ce,R=c(),V=M("p"),V.textContent=oe,H=c(),_=M("div"),_.innerHTML=me,D=c(),N=M("p"),N.innerHTML=pe,P=c(),g(z.$$.fragment),K=c(),g(E.$$.fragment),O=c(),F=M("p"),F.innerHTML=ue,ee=c(),Z=M("div"),Z.innerHTML=be,se=c(),Q=M("p"),Q.innerHTML=de,te=c(),w.c(),G=c(),v=M("blockquote"),v.innerHTML=fe,le=c(),g(B.$$.fragment),ne=c(),A=M("p"),this.h()},l(e){const l=_e("svelte-u9bgzb",document.head);n=y(l,"META",{name:!0,content:!0}),l.forEach(a),i=o(e),t=y(e,"P",{}),Te(t).forEach(a),d=o(e),J(u.$$.fragment,e),m=o(e),J(b.$$.fragment,e),s=o(e),J(p.$$.fragment,e),S=o(e),U.l(e),q=o(e),W=y(e,"P",{"data-svelte-h":!0}),C(W)!=="svelte-14ptojn"&&(W.innerHTML=ce),R=o(e),V=y(e,"P",{"data-svelte-h":!0}),C(V)!=="svelte-1263zlh"&&(V.textContent=oe),H=o(e),_=y(e,"DIV",{class:!0,"data-svelte-h":!0}),C(_)!=="svelte-4k6hvy"&&(_.innerHTML=me),D=o(e),N=y(e,"P",{"data-svelte-h":!0}),C(N)!=="svelte-1h6aycb"&&(N.innerHTML=pe),P=o(e),J(z.$$.fragment,e),K=o(e),J(E.$$.fragment,e),O=o(e),F=y(e,"P",{"data-svelte-h":!0}),C(F)!=="svelte-1o379vp"&&(F.innerHTML=ue),ee=o(e),Z=y(e,"DIV",{class:!0,"data-svelte-h":!0}),C(Z)!=="svelte-1cl7xbt"&&(Z.innerHTML=be),se=o(e),Q=y(e,"P",{"data-svelte-h":!0}),C(Q)!=="svelte-1x0kbto"&&(Q.innerHTML=de),te=o(e),w.l(e),G=o(e),v=y(e,"BLOCKQUOTE",{class:!0,"data-svelte-h":!0}),C(v)!=="svelte-5zecpw"&&(v.innerHTML=fe),le=o(e),J(B.$$.fragment,e),ne=o(e),A=y(e,"P",{}),Te(A).forEach(a),this.h()},h(){X(n,"name","hf:doc:metadata"),X(n,"content",Ie),X(_,"class","flex justify-center"),X(Z,"class","flex justify-center"),X(v,"class","tip")},m(e,l){Ze(document.head,n),r(e,i,l),r(e,t,l),r(e,d,l),T(u,e,l),r(e,m,l),T(b,e,l),r(e,s,l),T(p,e,l),r(e,S,l),I[x].m(e,l),r(e,q,l),r(e,W,l),r(e,R,l),r(e,V,l),r(e,H,l),r(e,_,l),r(e,D,l),r(e,N,l),r(e,P,l),T(z,e,l),r(e,K,l),T(E,e,l),r(e,O,l),r(e,F,l),r(e,ee,l),r(e,Z,l),r(e,se,l),r(e,Q,l),r(e,te,l),Y[$].m(e,l),r(e,G,l),r(e,v,l),r(e,le,l),T(B,e,l),r(e,ne,l),r(e,A,l),ae=!0},p(e,[l]){const Je={};l&1&&(Je.fw=e[0]),u.$set(Je);let re=x;x=Me(e),x!==re&&(je(),f(I[re],1,1,()=>{I[re]=null}),ke(),U=I[x],U\|\|(U=I[x]=he[x](e),U.c()),h(U,1),U.m(q.parentNode,q));let ie=$;$=ge(e),$!==ie&&(je(),f(Y[ie],1,1,()=>{Y[ie]=null}),ke(),w=Y[$],w\|\|(w=Y[$]=ye[$](e),w.c()),h(w,1),w.m(G.parentNode,G))},i(e){ae\|\|(h(u.$$.fragment,e),h(b.$$.fragment,e),h(p.$$.fragment,e),h(U),h(z.$$.fragment,e),h(E.$$.fragment,e),h(w),h(B.$$.fragment,e),ae=!0)},o(e){f(u.$$.fragment,e),f(b.$$.fragment,e),f(p.$$.fragment,e),f(U),f(z.$$.fragment,e),f(E.$$.fragment,e),f(w),f(B.$$.fragment,e),ae=!1},d(e){e&&(a(i),a(t),a(d),a(m),a(s),a(S),a(q),a(W),a(R),a(V),a(H),a(_),a(D),a(N),a(P),a(K),a(O),a(F),a(ee),a(Z),a(se),a(Q),a(te),a(G),a(v),a(le),a(ne),a(A)),a(n),k(u,e),k(b,e),k(p,e),I[x].d(e),k(z,e),k(E,e),Y[$].d(e),k(B,e)}}}const Ie='{"title":"Utilisation de modèles pré-entraînés","local":"utilisation-de-modèles-pré-entraînés","sections":[],"depth":1}';function Ye(j,n,i){let t="pt";return $e(()=>{const d=new URLSearchParams(window.location.search);i(0,t=d.get("fw")\|\|"pt")}),[t]}class Re extends we{constructor(n){super(),Ce(this,n,Ye,Be,Ue,{})}}export{Re as component};

Xet Storage Details

Size:: 17.2 kB
Xet hash:: 25bb258aa6e7e09ca7a6dd9184b845fb023886743adbe0152a7104bac06adfa7

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.