Buckets:

download
raw
18.3 kB
<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Fine-tune a model&quot;,&quot;local&quot;:&quot;fine-tune-a-model&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Train an adapter&quot;,&quot;local&quot;:&quot;train-an-adapter&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Save the adapter&quot;,&quot;local&quot;:&quot;save-the-adapter&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Quantization-aware training&quot;,&quot;local&quot;:&quot;quantization-aware-training&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/>
<link href="/docs/transformers/main/en/_app/immutable/entry/start.DYIO-UII.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/DSYYoqAA.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/BovLmb3z.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/entry/app.D2_KizbC.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/WyYpyc-L.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/DYBfcveH.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/DOLBYqcr.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/DsnmJJEf.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/nodes/0.ByO_lNbj.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/ZtBhN-Ei.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/nodes/708.D7ZPq8Iq.js" rel="modulepreload">
<link href="/docs/transformers/main/en/_app/immutable/chunks/D-dvW--N.js" rel="modulepreload">
<!--sv784v--><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;Fine-tune a model&quot;,&quot;local&quot;:&quot;fine-tune-a-model&quot;,&quot;sections&quot;:[{&quot;title&quot;:&quot;Train an adapter&quot;,&quot;local&quot;:&quot;train-an-adapter&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Save the adapter&quot;,&quot;local&quot;:&quot;save-the-adapter&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2},{&quot;title&quot;:&quot;Quantization-aware training&quot;,&quot;local&quot;:&quot;quantization-aware-training&quot;,&quot;sections&quot;:[],&quot;depth&quot;:2}],&quot;depth&quot;:1}"/><!---->
<link href="/docs/transformers/main/en/_app/immutable/assets/0.tn0RQdqM.css" rel="modulepreload"> <!--[--><!--[0--><!--[--><!--[0--><!--[--><p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: inline-end; margin-inline-start: 10px; display: inline-flex; position: relative; z-index: 10;"><div dir="ltr" class="inline-flex rounded-md max-sm:rounded-sm"><button class="inline-flex items-center gap-1 h-7 max-sm:h-7 px-2 max-sm:px-1.5 text-sm font-medium text-gray-800 border border-r-0 rounded-l-md max-sm:rounded-l-sm border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-live="polite"><span class="inline-flex items-center justify-center rounded-md p-0.5 max-sm:p-0 hover:text-gray-800 dark:hover:text-gray-200"><svg class="sm:size-3.5 size-3" 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><!----></span> <span dir="auto">Copy page</span></button> <button class="inline-flex items-center justify-center w-6 max-sm:w-5 h-7 max-sm:h-7 disabled:pointer-events-none text-sm text-gray-500 hover:text-gray-700 dark:hover:text-white rounded-r-md max-sm:rounded-r-sm border border-l transition border-gray-200 bg-white hover:shadow-inner dark:border-gray-850 dark:bg-gray-950 dark:text-gray-200 dark:hover:bg-gray-800" aria-haspopup="menu" aria-expanded="false" aria-label="Open copy menu"><svg class="transition-transform text-gray-400 overflow-visible sm:size-3.5 size-3 rotate-0" width="1em" height="1em" viewBox="0 0 12 7" fill="none" xmlns="http://www.w3.org/2000/svg"><path d="M1 1L6 6L11 1" stroke="currentColor"></path></svg><!----></button></div> <!--[-1--><!--]--></div><!----> <!--[0--><h1 class="relative group"><a id="fine-tune-a-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="#fine-tune-a-model"><span><svg 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><!---->Fine-tune a model<!----></span></h1><!--]--><!----> <p>Quantized weights can’t be updated by gradients directly, so you fine-tune a quantized model by training a small adapter on top of the frozen quantized weights. Training a <a href="https://huggingface.co/docs/peft/main/en/conceptual_guides/lora" rel="nofollow">LoRA</a> adapter on a 4-bit model is known as QLoRA, and it lets you fine-tune a model that wouldn’t fit in memory in bf16.</p> <p><a href="/docs/transformers/pr_49012/en/main_classes/trainer#transformers.Trainer">Trainer</a> checks the model before training to make sure:</p> <ul><li>The quantization method supports training. See the Fine-tune column in <a href="./selecting">Choose a method</a>.</li> <li>The model has a trainable adapter attached. Methods that support quantization-aware training are the exception, covered <a href="#quantization-aware-training">below</a>.</li> <li>The model isn’t compiled with <code>torch.compile</code>.</li></ul> <!--[1--><h2 class="relative group"><a id="train-an-adapter" 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="#train-an-adapter"><span><svg 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><!---->Train an adapter<!----></span></h2><!--]--><!----> <p>Install <a href="https://huggingface.co/docs/peft/index" rel="nofollow">PEFT</a>.</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 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="language-bash "><!---->pip install -U peft<!----></pre></div><!----> <p>Load the model with a method that supports training, attach a LoRA adapter with <a href="/docs/transformers/pr_49012/en/main_classes/peft#transformers.integrations.PeftAdapterMixin.add_adapter">add_adapter()</a>, and pass the model to <a href="/docs/transformers/pr_49012/en/main_classes/trainer#transformers.Trainer">Trainer</a>. The example below uses 4-bit <a href="./bitsandbytes">bitsandbytes</a> and assumes <code>train_dataset</code> is a tokenized dataset.</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 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="language-py "><!----><span class="hljs-keyword">import</span> torch
<span class="hljs-keyword">from</span> peft <span class="hljs-keyword">import</span> LoraConfig
<span class="hljs-keyword">from</span> transformers <span class="hljs-keyword">import</span> AutoModelForCausalLM, BitsAndBytesConfig, Trainer, TrainingArguments
quantization_config = BitsAndBytesConfig(
load_in_4bit=<span class="hljs-literal">True</span>,
bnb_4bit_quant_type=<span class="hljs-string">&quot;nf4&quot;</span>,
bnb_4bit_compute_dtype=torch.bfloat16,
)
model = AutoModelForCausalLM.from_pretrained(
<span class="hljs-string">&quot;Qwen/Qwen3-0.6B&quot;</span>,
quantization_config=quantization_config,
device_map=<span class="hljs-string">&quot;auto&quot;</span>,
)
lora_config = LoraConfig(r=<span class="hljs-number">16</span>, lora_alpha=<span class="hljs-number">32</span>, target_modules=<span class="hljs-string">&quot;all-linear&quot;</span>, task_type=<span class="hljs-string">&quot;CAUSAL_LM&quot;</span>)
model.add_adapter(lora_config)
trainer = Trainer(
model=model,
args=TrainingArguments(output_dir=<span class="hljs-string">&quot;qwen3-qlora&quot;</span>, bf16=<span class="hljs-literal">True</span>),
train_dataset=train_dataset,
)
trainer.train()<!----></pre></div><!----> <p>Only the adapter weights train. The quantized base weights stay frozen.</p> <!--[1--><h2 class="relative group"><a id="save-the-adapter" 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="#save-the-adapter"><span><svg 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><!---->Save the adapter<!----></span></h2><!--]--><!----> <p><a href="/docs/transformers/pr_49012/en/main_classes/model#transformers.PreTrainedModel.save_pretrained">save_pretrained()</a> saves only the adapter weights, which are a few megabytes. This works with every method, including ones that can’t save their quantized weights.</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 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="language-py "><!---->model.save_pretrained(<span class="hljs-string">&quot;qwen3-qlora-adapter&quot;</span>)<!----></pre></div><!----> <p>To use the adapter, load the base model with the same quantization config and load the adapter on top of it. See <a href="../peft#loading-an-adapter">Loading an adapter</a> for details.</p> <!--[1--><h2 class="relative group"><a id="quantization-aware-training" 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="#quantization-aware-training"><span><svg 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><!---->Quantization-aware training<!----></span></h2><!--]--><!----> <p><a href="./bitnet">BitNet</a> and <a href="./compressed_tensors">compressed-tensors</a> can train the full model without an adapter, as long as the weights stay in a trainable form.</p> <ul><li>BitNet needs <code>linear_class="autobitlinear"</code> and <code>quantization_mode="online"</code> in <a href="/docs/transformers/pr_49012/en/main_classes/quantization#transformers.BitNetQuantConfig">BitNetQuantConfig</a>.</li> <li>compressed-tensors needs a checkpoint that isn’t compressed, or <code>dequantize=True</code> in <a href="/docs/transformers/pr_49012/en/main_classes/quantization#transformers.CompressedTensorsConfig">CompressedTensorsConfig</a>. Checkpoints that use the FP8 kernels can’t train.</li></ul> <p>See <a href="./concept_guide">Quantization concepts</a> for how quantization-aware training differs from post-training quantization.</p> <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/quantization/finetune.md" target="_blank"><svg class="mr-1" 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="M31,16l-7,7l-1.41-1.41L28.17,16l-5.58-5.59L24,9l7,7z"></path><path d="M1,16l7-7l1.41,1.41L3.83,16l5.58,5.59L8,23l-7-7z"></path><path d="M12.419,25.484L17.639,6.552l1.932,0.518L14.351,26.002z"></path></svg><!----> <span><span class="underline">Update</span> on GitHub</span></a><!----> <p></p><!--]--><!----><!--]--><!--]--><!--]--> <!--[-1--><!--]--><!--]-->
<script>
{
__sveltekit_1m6yd5j = {
base: "/docs/transformers/main/en",
assets: "/docs/transformers/main/en"
};
const element = document.currentScript.parentElement;
Promise.all([
import("/docs/transformers/main/en/_app/immutable/entry/start.DYIO-UII.js"),
import("/docs/transformers/main/en/_app/immutable/entry/app.D2_KizbC.js")
]).then(([kit, app]) => {
kit.start(app, element, {
node_ids: [0, 708],
data: [null,null],
form: null,
error: null
});
});
}
</script>

Xet Storage Details

Size:
18.3 kB
·
Xet hash:
b7a6e959c1034db7c12301b15e6a6b0ed43cd16292b3cc908a90d98a7355c285

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