Hugging Face's logo

Buckets:
hf-doc-build
/
doc-dev

Files
xet
hf-doc-build/doc-dev / transformers /main /en /troubleshooting.html
rtrm's picture
rtrm
download
raw
46.9 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Troubleshoot&quot;,&quot;local&quot;:&quot;troubleshoot&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Firewalled environments&quot;,&quot;local&quot;:&quot;firewalled-environments&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;CUDA out of memory&quot;,&quot;local&quot;:&quot;cuda-out-of-memory&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Unable to load a saved TensorFlow model&quot;,&quot;local&quot;:&quot;unable-to-load-a-saved-tensorflow-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;ImportError&quot;,&quot;local&quot;:&quot;importerror&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;CUDA error: device-side assert triggered&quot;,&quot;local&quot;:&quot;cuda-error-device-side-assert-triggered&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Incorrect output when padding tokens aren’t masked&quot;,&quot;local&quot;:&quot;incorrect-output-when-padding-tokens-arent-masked&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;ValueError: Unrecognized configuration class XYZ for this kind of AutoModel&quot;,&quot;local&quot;:&quot;valueerror-unrecognized-configuration-class-xyz-for-this-kind-of-automodel&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}">
 <link href="/docs/transformers/main/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/entry/start.2135b7e6.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/scheduler.25b97de1.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/singletons.0f2b7d5f.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/index.e188933d.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/paths.3d04d2c6.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/entry/app.24372c84.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/index.d9030fc9.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/nodes/0.026d2fdd.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/each.e59479a4.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/nodes/438.01822960.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/Tip.baa67368.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/Youtube.eaf1a617.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/CodeBlock.e6cd0d95.js">
 <link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/EditOnGithub.91d95064.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Troubleshoot&quot;,&quot;local&quot;:&quot;troubleshoot&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Firewalled environments&quot;,&quot;local&quot;:&quot;firewalled-environments&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;CUDA out of memory&quot;,&quot;local&quot;:&quot;cuda-out-of-memory&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Unable to load a saved TensorFlow model&quot;,&quot;local&quot;:&quot;unable-to-load-a-saved-tensorflow-model&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;ImportError&quot;,&quot;local&quot;:&quot;importerror&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;CUDA error: device-side assert triggered&quot;,&quot;local&quot;:&quot;cuda-error-device-side-assert-triggered&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Incorrect output when padding tokens aren’t masked&quot;,&quot;local&quot;:&quot;incorrect-output-when-padding-tokens-arent-masked&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;ValueError: Unrecognized configuration class XYZ for this kind of AutoModel&quot;,&quot;local&quot;:&quot;valueerror-unrecognized-configuration-class-xyz-for-this-kind-of-automodel&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <h1 class="relative group"><a id="troubleshoot" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#troubleshoot"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Troubleshoot</span></h1> <p data-svelte-h="svelte-1cx76wd">Sometimes errors occur, but we are here to help! This guide covers some of the most common issues we’ve seen and how you can resolve them. However, this guide isn’t meant to be a comprehensive collection of every 🤗 Transformers issue. For more help with troubleshooting your issue, try:</p> <iframe class="w-full xl:w-4/6 h-80" src="https://www.youtube-nocookie.com/embed/S2EEG3JIt2A" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe> <ol data-svelte-h="svelte-wyi91t"><li>Asking for help on the <a href="https://discuss.huggingface.co/" rel="nofollow">forums</a>. There are specific categories you can post your question to, like <a href="https://discuss.huggingface.co/c/beginners/5" rel="nofollow">Beginners</a> or <a href="https://discuss.huggingface.co/c/transformers/9" rel="nofollow">🤗 Transformers</a>. Make sure you write a good descriptive forum post with some reproducible code to maximize the likelihood that your problem is solved!</li></ol> <iframe class="w-full xl:w-4/6 h-80" src="https://www.youtube-nocookie.com/embed/_PAli-V4wj0" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe> <ol start="2" data-svelte-h="svelte-l1kxjh"><li><p>Create an <a href="https://github.com/huggingface/transformers/issues/new/choose" rel="nofollow">Issue</a> on the 🤗 Transformers repository if it is a bug related to the library. Try to include as much information describing the bug as possible to help us better figure out what’s wrong and how we can fix it.</p></li> <li><p>Check the <a href="migration">Migration</a> guide if you use an older version of 🤗 Transformers since some important changes have been introduced between versions.</p></li></ol> <p data-svelte-h="svelte-c2f5jg">For more details about troubleshooting and getting help, take a look at <a href="https://huggingface.co/course/chapter8/1?fw=pt" rel="nofollow">Chapter 8</a> of the Hugging Face course.</p> <h2 class="relative group"><a id="firewalled-environments" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#firewalled-environments"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Firewalled environments</span></h2> <p data-svelte-h="svelte-lpgdhl">Some GPU instances on cloud and intranet setups are firewalled to external connections, resulting in a connection error. When your script attempts to download model weights or datasets, the download will hang and then timeout with the following message:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->ValueError: Connection error, <span class="hljs-built_in">and</span> we cannot <span class="hljs-keyword">find</span> the requested <span class="hljs-keyword">files</span> in the cached path.
Please <span class="hljs-keyword">try</span> again <span class="hljs-built_in">or</span> <span class="hljs-keyword">make</span> sure your Internet connection <span class="hljs-keyword">is</span> <span class="hljs-keyword">on</span>.<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-170em04">In this case, you should try to run 🤗 Transformers on <a href="installation#offline-mode">offline mode</a> to avoid the connection error.</p> <h2 class="relative group"><a id="cuda-out-of-memory" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#cuda-out-of-memory"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>CUDA out of memory</span></h2> <p data-svelte-h="svelte-28duqs">Training large models with millions of parameters can be challenging without the appropriate hardware. A common error you may encounter when the GPU runs out of memory is:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-attribute">CUDA</span> out of memory. Tried to allocate <span class="hljs-number">256</span>.<span class="hljs-number">00</span> MiB (GPU <span class="hljs-number">0</span>; <span class="hljs-number">11</span>.<span class="hljs-number">17</span> GiB total capacity; <span class="hljs-number">9</span>.<span class="hljs-number">70</span> GiB already allocated; <span class="hljs-number">179</span>.<span class="hljs-number">81</span> MiB free; <span class="hljs-number">9</span>.<span class="hljs-number">85</span> GiB reserved in total by PyTorch)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1wd6425">Here are some potential solutions you can try to lessen memory use:</p> <ul data-svelte-h="svelte-3xmtjw"><li>Reduce the <a href="main_classes/trainer#transformers.TrainingArguments.per_device_train_batch_size"><code>per_device_train_batch_size</code></a> value in <a href="/docs/transformers/main/en/main_classes/trainer#transformers.TrainingArguments">TrainingArguments</a>.</li> <li>Try using <a href="main_classes/trainer#transformers.TrainingArguments.gradient_accumulation_steps"><code>gradient_accumulation_steps</code></a> in <a href="/docs/transformers/main/en/main_classes/trainer#transformers.TrainingArguments">TrainingArguments</a> to effectively increase overall batch size.</li></ul> <div class="course-tip bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p data-svelte-h="svelte-xgjmzv">Refer to the Performance <a href="performance">guide</a> for more details about memory-saving techniques.</p></div> <h2 class="relative group"><a id="unable-to-load-a-saved-tensorflow-model" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#unable-to-load-a-saved-tensorflow-model"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Unable to load a saved TensorFlow model</span></h2> <p data-svelte-h="svelte-dxswyr">TensorFlow’s <a href="https://www.tensorflow.org/tutorials/keras/save_and_load#save_the_entire_model" rel="nofollow">model.save</a> method will save the entire model - architecture, weights, training configuration - in a single file. However, when you load the model file again, you may run into an error because 🤗 Transformers may not load all the TensorFlow-related objects in the model file. To avoid issues with saving and loading TensorFlow models, we recommend you:</p> <ul data-svelte-h="svelte-axsa61"><li>Save the model weights as a <code>h5</code> file extension with <a href="https://www.tensorflow.org/tutorials/keras/save_and_load#save_the_entire_model" rel="nofollow"><code>model.save_weights</code></a> and then reload the model with <a href="/docs/transformers/main/en/main_classes/model#transformers.TFPreTrainedModel.from_pretrained">from_pretrained()</a>:</li></ul> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> TFPreTrainedModel
<span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">from</span> tensorflow <span class="hljs-keyword">import</span> keras
<span class="hljs-meta">&gt;&gt;&gt; </span>model.save_weights(<span class="hljs-string">&quot;some_folder/tf_model.h5&quot;</span>)
<span class="hljs-meta">&gt;&gt;&gt; </span>model = TFPreTrainedModel.from_pretrained(<span class="hljs-string">&quot;some_folder&quot;</span>)<!-- HTML_TAG_END --></pre></div> <ul data-svelte-h="svelte-m1h2rm"><li>Save the model with <code>~TFPretrainedModel.save_pretrained</code> and load it again with <a href="/docs/transformers/main/en/main_classes/model#transformers.TFPreTrainedModel.from_pretrained">from_pretrained()</a>:</li></ul> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> TFPreTrainedModel
<span class="hljs-meta">&gt;&gt;&gt; </span>model.save_pretrained(<span class="hljs-string">&quot;path_to/model&quot;</span>)
<span class="hljs-meta">&gt;&gt;&gt; </span>model = TFPreTrainedModel.from_pretrained(<span class="hljs-string">&quot;path_to/model&quot;</span>)<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="importerror" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#importerror"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>ImportError</span></h2> <p data-svelte-h="svelte-1yk3oka">Another common error you may encounter, especially if it is a newly released model, is <code>ImportError</code>:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->ImportError: cannot <span class="hljs-keyword">import</span> <span class="hljs-type">name</span> <span class="hljs-string">&#x27;ImageGPTImageProcessor&#x27;</span> <span class="hljs-keyword">from</span> <span class="hljs-string">&#x27;transformers&#x27;</span> (<span class="hljs-type">unknown</span> <span class="hljs-keyword">location</span>)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-4f4dig">For these error types, check to make sure you have the latest version of 🤗 Transformers installed to access the most recent models:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->pip install transformers --upgrade<!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="cuda-error-device-side-assert-triggered" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#cuda-error-device-side-assert-triggered"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>CUDA error: device-side assert triggered</span></h2> <p data-svelte-h="svelte-1u0kwcf">Sometimes you may run into a generic CUDA error about an error in the device code.</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START -->RuntimeError: CUDA <span class="hljs-literal">error</span>: device-<span class="hljs-literal">side</span> <span class="hljs-keyword">assert</span> triggered<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-aeboz0">You should try to run the code on a CPU first to get a more descriptive error message. Add the following environment variable to the beginning of your code to switch to a CPU:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">import</span> os
<span class="hljs-meta">&gt;&gt;&gt; </span>os.environ[<span class="hljs-string">&quot;CUDA_VISIBLE_DEVICES&quot;</span>] = <span class="hljs-string">&quot;&quot;</span><!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-ctfxbh">Another option is to get a better traceback from the GPU. Add the following environment variable to the beginning of your code to get the traceback to point to the source of the error:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">import</span> os
<span class="hljs-meta">&gt;&gt;&gt; </span>os.environ[<span class="hljs-string">&quot;CUDA_LAUNCH_BLOCKING&quot;</span>] = <span class="hljs-string">&quot;1&quot;</span><!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="incorrect-output-when-padding-tokens-arent-masked" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#incorrect-output-when-padding-tokens-arent-masked"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>Incorrect output when padding tokens aren’t masked</span></h2> <p data-svelte-h="svelte-1hu02cc">In some cases, the output <code>hidden_state</code> may be incorrect if the <code>input_ids</code> include padding tokens. To demonstrate, load a model and tokenizer. You can access a model’s <code>pad_token_id</code> to see its value. The <code>pad_token_id</code> may be <code>None</code> for some models, but you can always manually set it.</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoModelForSequenceClassification
<span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">import</span> torch
<span class="hljs-meta">&gt;&gt;&gt; </span>model = AutoModelForSequenceClassification.from_pretrained(<span class="hljs-string">&quot;google-bert/bert-base-uncased&quot;</span>)
<span class="hljs-meta">&gt;&gt;&gt; </span>model.config.pad_token_id
<span class="hljs-number">0</span><!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-i1hsji">The following example shows the output without masking the padding tokens:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span>input_ids = torch.tensor([[<span class="hljs-number">7592</span>, <span class="hljs-number">2057</span>, <span class="hljs-number">2097</span>, <span class="hljs-number">2393</span>, <span class="hljs-number">9611</span>, <span class="hljs-number">2115</span>], [<span class="hljs-number">7592</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>]])
<span class="hljs-meta">&gt;&gt;&gt; </span>output = model(input_ids)
<span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-built_in">print</span>(output.logits)
tensor([[ <span class="hljs-number">0.0082</span>, -<span class="hljs-number">0.2307</span>],
 [ <span class="hljs-number">0.1317</span>, -<span class="hljs-number">0.1683</span>]], grad_fn=&lt;AddmmBackward0&gt;)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-101z1g5">Here is the actual output of the second sequence:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span>input_ids = torch.tensor([[<span class="hljs-number">7592</span>]])
<span class="hljs-meta">&gt;&gt;&gt; </span>output = model(input_ids)
<span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-built_in">print</span>(output.logits)
tensor([[-<span class="hljs-number">0.1008</span>, -<span class="hljs-number">0.4061</span>]], grad_fn=&lt;AddmmBackward0&gt;)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1p57h5c">Most of the time, you should provide an <code>attention_mask</code> to your model to ignore the padding tokens to avoid this silent error. Now the output of the second sequence matches its actual output:</p> <div class="course-tip bg-gradient-to-br dark:bg-gradient-to-r before:border-green-500 dark:before:border-green-800 from-green-50 dark:from-gray-900 to-white dark:to-gray-950 border border-green-50 text-green-700 dark:text-gray-400"><p data-svelte-h="svelte-fvd42m">By default, the tokenizer creates an <code>attention_mask</code> for you based on your specific tokenizer’s defaults.</p></div> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span>attention_mask = torch.tensor([[<span class="hljs-number">1</span>, <span class="hljs-number">1</span>, <span class="hljs-number">1</span>, <span class="hljs-number">1</span>, <span class="hljs-number">1</span>, <span class="hljs-number">1</span>], [<span class="hljs-number">1</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>, <span class="hljs-number">0</span>]])
<span class="hljs-meta">&gt;&gt;&gt; </span>output = model(input_ids, attention_mask=attention_mask)
<span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-built_in">print</span>(output.logits)
tensor([[ <span class="hljs-number">0.0082</span>, -<span class="hljs-number">0.2307</span>],
 [-<span class="hljs-number">0.1008</span>, -<span class="hljs-number">0.4061</span>]], grad_fn=&lt;AddmmBackward0&gt;)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-150uvc6">🤗 Transformers doesn’t automatically create an <code>attention_mask</code> to mask a padding token if it is provided because:</p> <ul data-svelte-h="svelte-qm9a5i"><li>Some models don’t have a padding token.</li> <li>For some use-cases, users want a model to attend to a padding token.</li></ul> <h2 class="relative group"><a id="valueerror-unrecognized-configuration-class-xyz-for-this-kind-of-automodel" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#valueerror-unrecognized-configuration-class-xyz-for-this-kind-of-automodel"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>ValueError: Unrecognized configuration class XYZ for this kind of AutoModel</span></h2> <p data-svelte-h="svelte-11l04n2">Generally, we recommend using the <a href="/docs/transformers/main/en/model_doc/auto#transformers.AutoModel">AutoModel</a> class to load pretrained instances of models. This class
can automatically infer and load the correct architecture from a given checkpoint based on the configuration. If you see
this <code>ValueError</code> when loading a model from a checkpoint, this means the Auto class couldn’t find a mapping from
the configuration in the given checkpoint to the kind of model you are trying to load. Most commonly, this happens when a
checkpoint doesn’t support a given task.
For instance, you’ll see this error in the following example because there is no GPT2 for question answering:</p> <div class="code-block relative"><div class="absolute top-2.5 right-4"><button class="inline-flex items-center relative text-sm focus:text-green-500 cursor-pointer focus:outline-none transition duration-200 ease-in-out opacity-0 mx-0.5 text-gray-600 " title="code excerpt" type="button"><svg class="" xmlns="http://www.w3.org/2000/svg" aria-hidden="true" fill="currentColor" focusable="false" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 32 32"><path d="M28,10V28H10V10H28m0-2H10a2,2,0,0,0-2,2V28a2,2,0,0,0,2,2H28a2,2,0,0,0,2-2V10a2,2,0,0,0-2-2Z" transform="translate(0)"></path><path d="M4,18H2V4A2,2,0,0,1,4,2H18V4H4Z" transform="translate(0)"></path><rect fill="none" width="32" height="32"></rect></svg> <div class="absolute pointer-events-none transition-opacity bg-black text-white py-1 px-2 leading-tight rounded font-normal shadow left-1/2 top-full transform -translate-x-1/2 translate-y-2 opacity-0"><div class="absolute bottom-full left-1/2 transform -translate-x-1/2 w-0 h-0 border-black border-4 border-t-0" style="border-left-color: transparent; border-right-color: transparent; "></div> Copied</div></button></div> <pre class=""><!-- HTML_TAG_START --><span class="hljs-meta">&gt;&gt;&gt; </span><span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoProcessor, AutoModelForQuestionAnswering
<span class="hljs-meta">&gt;&gt;&gt; </span>processor = AutoProcessor.from_pretrained(<span class="hljs-string">&quot;openai-community/gpt2-medium&quot;</span>)
<span class="hljs-meta">&gt;&gt;&gt; </span>model = AutoModelForQuestionAnswering.from_pretrained(<span class="hljs-string">&quot;openai-community/gpt2-medium&quot;</span>)
ValueError: Unrecognized configuration <span class="hljs-keyword">class</span> &lt;<span class="hljs-keyword">class</span> <span class="hljs-string">&#x27;transformers.models.gpt2.configuration_gpt2.GPT2Config&#x27;</span>&gt; <span class="hljs-keyword">for</span> this kind of AutoModel: AutoModelForQuestionAnswering.
Model <span class="hljs-built_in">type</span> should be one of AlbertConfig, BartConfig, BertConfig, BigBirdConfig, BigBirdPegasusConfig, BloomConfig, ...<!-- HTML_TAG_END --></pre></div> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/transformers/blob/main/docs/source/en/troubleshooting.md" target="_blank"><span data-svelte-h="svelte-1kd6by1">&lt;</span> <span data-svelte-h="svelte-x0xyl0">&gt;</span> <span data-svelte-h="svelte-1dajgef"><span class="underline ml-1.5">Update</span> on GitHub</span></a> <p></p>
<script>
 {
 __sveltekit_1xexzbk = {
 assets: "/docs/transformers/main/en",
 base: "/docs/transformers/main/en",
 env: {}
 };
 const element = document.currentScript.parentElement;
 const data = [null,null];
 Promise.all([
 import("/docs/transformers/main/en/_app/immutable/entry/start.2135b7e6.js"),
 import("/docs/transformers/main/en/_app/immutable/entry/app.24372c84.js")
 ]).then(([kit, app]) => {
 kit.start(app, element, {
 node_ids: [0, 438],
 data,
 form: null,
 error: null
 });
 });
 }
 </script>

Xet Storage Details

Size:
46.9 kB
·
Xet hash:
636ba9966475db10028eed389c40664e0d6d26b399eae1f1925209bcca02123f

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