Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"GGUF and interaction with Transformers","local":"gguf-and-interaction-with-transformers","sections":[{"title":"Support within Transformers","local":"support-within-transformers","sections":[{"title":"Supported quantization types","local":"supported-quantization-types","sections":[],"depth":3},{"title":"Supported model architectures","local":"supported-model-architectures","sections":[],"depth":3}],"depth":2},{"title":"Example usage","local":"example-usage","sections":[],"depth":2}],"depth":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/23.dea0f42a.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="{"title":"GGUF and interaction with Transformers","local":"gguf-and-interaction-with-transformers","sections":[{"title":"Support within Transformers","local":"support-within-transformers","sections":[{"title":"Supported quantization types","local":"supported-quantization-types","sections":[],"depth":3},{"title":"Supported model architectures","local":"supported-model-architectures","sections":[],"depth":3}],"depth":2},{"title":"Example usage","local":"example-usage","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <h1 class="relative group"><a id="gguf-and-interaction-with-transformers" 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="#gguf-and-interaction-with-transformers"><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>GGUF and interaction with Transformers</span></h1> <p data-svelte-h="svelte-117klwz">The GGUF file format is used to store models for inference with <a href="https://github.com/ggerganov/ggml" rel="nofollow">GGML</a> and other | |
| libraries that depend on it, like the very popular <a href="https://github.com/ggerganov/llama.cpp" rel="nofollow">llama.cpp</a> or | |
| <a href="https://github.com/ggerganov/whisper.cpp" rel="nofollow">whisper.cpp</a>.</p> <p data-svelte-h="svelte-14alsnt">It is a file format <a href="https://huggingface.co/docs/hub/en/gguf" rel="nofollow">supported by the Hugging Face Hub</a> with features | |
| allowing for quick inspection of tensors and metadata within the file.</p> <p data-svelte-h="svelte-1lcntb8">This file format is designed as a “single-file-format” where a single file usually contains both the configuration | |
| attributes, the tokenizer vocabulary and other attributes, as well as all tensors to be loaded in the model. These | |
| files come in different formats according to the quantization type of the file. We briefly go over some of them | |
| <a href="https://huggingface.co/docs/hub/en/gguf#quantization-types" rel="nofollow">here</a>.</p> <h2 class="relative group"><a id="support-within-transformers" 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="#support-within-transformers"><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>Support within Transformers</span></h2> <p data-svelte-h="svelte-1fcaojo">We have added the ability to load <code>gguf</code> files within <code>transformers</code> in order to offer further training/fine-tuning | |
| capabilities to gguf models, before converting back those models to <code>gguf</code> to use within the <code>ggml</code> ecosystem. When | |
| loading a model, we first dequantize it to fp32, before loading the weights to be used in PyTorch.</p> <blockquote data-svelte-h="svelte-13g437o"><p>[!NOTE] | |
| The support is still very exploratory and we welcome contributions in order to solidify it across quantization types | |
| and model architectures.</p></blockquote> <p data-svelte-h="svelte-lue5cc">For now, here are the supported model architectures and quantization types:</p> <h3 class="relative group"><a id="supported-quantization-types" 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="#supported-quantization-types"><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>Supported quantization types</span></h3> <p data-svelte-h="svelte-1r2kpez">The initial supported quantization types are decided according to the popular quantized files that have been shared | |
| on the Hub.</p> <ul data-svelte-h="svelte-1f29qgp"><li>F32</li> <li>F16</li> <li>BF16</li> <li>Q4_0</li> <li>Q4_1</li> <li>Q5_0</li> <li>Q5_1</li> <li>Q8_0</li> <li>Q2_K</li> <li>Q3_K</li> <li>Q4_K</li> <li>Q5_K</li> <li>Q6_K</li> <li>IQ1_S</li> <li>IQ1_M</li> <li>IQ2_XXS</li> <li>IQ2_XS</li> <li>IQ2_S</li> <li>IQ3_XXS</li> <li>IQ3_S</li> <li>IQ4_XS</li> <li>IQ4_NL</li></ul> <blockquote data-svelte-h="svelte-tc3a5k"><p>[!NOTE] | |
| To support gguf dequantization, <code>gguf>=0.10.0</code> installation is required.</p></blockquote> <h3 class="relative group"><a id="supported-model-architectures" 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="#supported-model-architectures"><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>Supported model architectures</span></h3> <p data-svelte-h="svelte-tmyftl">For now the supported model architectures are the architectures that have been very popular on the Hub, namely:</p> <ul data-svelte-h="svelte-xxd66r"><li>LLaMa</li> <li>Mistral</li> <li>Qwen2</li> <li>Qwen2Moe</li> <li>Phi3</li></ul> <h2 class="relative group"><a id="example-usage" 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="#example-usage"><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>Example usage</span></h2> <p data-svelte-h="svelte-1lj1hm3">In order to load <code>gguf</code> files in <code>transformers</code>, you should specify the <code>gguf_file</code> argument to the <code>from_pretrained</code> | |
| methods of both tokenizers and models. Here is how one would load a tokenizer and a model, which can be loaded | |
| from the exact same file:</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-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoTokenizer, AutoModelForCausalLM | |
| model_id = <span class="hljs-string">"TheBloke/TinyLlama-1.1B-Chat-v1.0-GGUF"</span> | |
| filename = <span class="hljs-string">"tinyllama-1.1b-chat-v1.0.Q6_K.gguf"</span> | |
| tokenizer = AutoTokenizer.from_pretrained(model_id, gguf_file=filename) | |
| model = AutoModelForCausalLM.from_pretrained(model_id, gguf_file=filename)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-16kc502">Now you have access to the full, unquantized version of the model in the PyTorch ecosystem, where you can combine it | |
| with a plethora of other tools.</p> <p data-svelte-h="svelte-uk9y6w">In order to convert back to a <code>gguf</code> file, we recommend using the | |
| <a href="https://github.com/ggerganov/llama.cpp/blob/master/convert-hf-to-gguf.py" rel="nofollow"><code>convert-hf-to-gguf.py</code> file</a> from llama.cpp.</p> <p data-svelte-h="svelte-eicl62">Here’s how you would complete the script above to save the model and export it back to <code>gguf</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 -->tokenizer.save_pretrained(<span class="hljs-string">'directory'</span>) | |
| model.save_pretrained(<span class="hljs-string">'directory'</span>) | |
| !python ${path_to_llama_cpp}/convert-hf-to-gguf.py ${directory}<!-- 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/gguf.md" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span data-svelte-h="svelte-x0xyl0">></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, 23], | |
| data, | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 16.5 kB
- Xet hash:
- a2b3e9b6be365bea9519814f5aac103e7d5e23cc2646480627d2b2ae628c55d2
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.