Buckets:
| <meta charset="utf-8" /><meta name="hf:doc:metadata" content="{"title":"DPO Trainer","local":"dpo-trainer","sections":[{"title":"Overview","local":"overview","sections":[],"depth":2},{"title":"Quick start","local":"quick-start","sections":[],"depth":2},{"title":"Expected dataset type and format","local":"expected-dataset-type-and-format","sections":[],"depth":2},{"title":"Looking deeper into the DPO method","local":"looking-deeper-into-the-dpo-method","sections":[{"title":"Preprocessing and tokenization","local":"preprocessing-and-tokenization","sections":[],"depth":3},{"title":"Computing the loss","local":"computing-the-loss","sections":[{"title":"Loss Types","local":"loss-types","sections":[],"depth":4}],"depth":3}],"depth":2},{"title":"Logged metrics","local":"logged-metrics","sections":[],"depth":2},{"title":"Customization","local":"customization","sections":[{"title":"Compatibility and constraints","local":"compatibility-and-constraints","sections":[],"depth":3},{"title":"Multi-loss combinations","local":"multi-loss-combinations","sections":[],"depth":3},{"title":"Model initialization","local":"model-initialization","sections":[],"depth":3},{"title":"Train adapters with PEFT","local":"train-adapters-with-peft","sections":[],"depth":3},{"title":"Train with Liger Kernel","local":"train-with-liger-kernel","sections":[],"depth":3},{"title":"Rapid Experimentation for DPO","local":"rapid-experimentation-for-dpo","sections":[],"depth":3},{"title":"Train with Unsloth","local":"train-with-unsloth","sections":[],"depth":3}],"depth":2},{"title":"Tool Calling with DPO","local":"tool-calling-with-dpo","sections":[],"depth":2},{"title":"Training Vision Language Models","local":"training-vision-language-models","sections":[],"depth":2},{"title":"DPOTrainer","local":"trl.DPOTrainer","sections":[],"depth":2},{"title":"DPOConfig","local":"trl.DPOConfig","sections":[],"depth":2}],"depth":1}"> | |
| <link href="/docs/trl/pr_5607/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/entry/start.151d81bd.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/scheduler.7b731bd4.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/singletons.2cf51804.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/index.ac28c20f.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/paths.ba01f37d.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/entry/app.3d9a91c0.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/preload-helper.e1689b3a.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/index.cc268345.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/nodes/0.cd288160.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/each.e59479a4.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/nodes/16.7e571f8e.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/MermaidChart.svelte_svelte_type_style_lang.f0d99f98.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/Docstring.03f7b462.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/CodeBlock.169a125f.js"> | |
| <link rel="modulepreload" href="/docs/trl/pr_5607/en/_app/immutable/chunks/ExampleCodeBlock.415f9452.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{"title":"DPO Trainer","local":"dpo-trainer","sections":[{"title":"Overview","local":"overview","sections":[],"depth":2},{"title":"Quick start","local":"quick-start","sections":[],"depth":2},{"title":"Expected dataset type and format","local":"expected-dataset-type-and-format","sections":[],"depth":2},{"title":"Looking deeper into the DPO method","local":"looking-deeper-into-the-dpo-method","sections":[{"title":"Preprocessing and tokenization","local":"preprocessing-and-tokenization","sections":[],"depth":3},{"title":"Computing the loss","local":"computing-the-loss","sections":[{"title":"Loss Types","local":"loss-types","sections":[],"depth":4}],"depth":3}],"depth":2},{"title":"Logged metrics","local":"logged-metrics","sections":[],"depth":2},{"title":"Customization","local":"customization","sections":[{"title":"Compatibility and constraints","local":"compatibility-and-constraints","sections":[],"depth":3},{"title":"Multi-loss combinations","local":"multi-loss-combinations","sections":[],"depth":3},{"title":"Model initialization","local":"model-initialization","sections":[],"depth":3},{"title":"Train adapters with PEFT","local":"train-adapters-with-peft","sections":[],"depth":3},{"title":"Train with Liger Kernel","local":"train-with-liger-kernel","sections":[],"depth":3},{"title":"Rapid Experimentation for DPO","local":"rapid-experimentation-for-dpo","sections":[],"depth":3},{"title":"Train with Unsloth","local":"train-with-unsloth","sections":[],"depth":3}],"depth":2},{"title":"Tool Calling with DPO","local":"tool-calling-with-dpo","sections":[],"depth":2},{"title":"Training Vision Language Models","local":"training-vision-language-models","sections":[],"depth":2},{"title":"DPOTrainer","local":"trl.DPOTrainer","sections":[],"depth":2},{"title":"DPOConfig","local":"trl.DPOConfig","sections":[],"depth":2}],"depth":1}"><!-- HEAD_svelte-u9bgzb_END --> <p></p> <div class="items-center shrink-0 min-w-[100px] max-sm:min-w-[50px] justify-end ml-auto flex" style="float: right; margin-left: 10px; display: inline-flex; position: relative; z-index: 10;"><div 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>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> </div> <h1 class="relative group"><a id="dpo-trainer" 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="#dpo-trainer"><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>DPO Trainer</span></h1> <p data-svelte-h="svelte-1tlxfr9"><a href="https://huggingface.co/models?other=dpo,trl" rel="nofollow"><img src="https://img.shields.io/badge/All_models-DPO-blue" alt="All_models-DPO-blue"></a> <a href="https://github.com/huggingface/smol-course/tree/main/2_preference_alignment" rel="nofollow"><img src="https://img.shields.io/badge/smol_course-Chapter_2-yellow" alt="smol_course-Chapter_2-yellow"></a></p> <h2 class="relative group"><a id="overview" 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="#overview"><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>Overview</span></h2> <p data-svelte-h="svelte-f8b3a9">TRL supports the Direct Preference Optimization (DPO) Trainer for training language models, as described in the paper <a href="https://huggingface.co/papers/2305.18290" rel="nofollow">Direct Preference Optimization: Your Language Model is Secretly a Reward Model</a> by <a href="https://huggingface.co/rmrafailov" rel="nofollow">Rafael Rafailov</a>, Archit Sharma, Eric Mitchell, <a href="https://huggingface.co/ermonste" rel="nofollow">Stefano Ermon</a>, <a href="https://huggingface.co/manning" rel="nofollow">Christopher D. Manning</a>, <a href="https://huggingface.co/cbfinn" rel="nofollow">Chelsea Finn</a>.</p> <p data-svelte-h="svelte-vfdo9a">The abstract from the paper is the following:</p> <blockquote data-svelte-h="svelte-79s4co"><p>While large-scale unsupervised language models (LMs) learn broad world knowledge and some reasoning skills, achieving precise control of their behavior is difficult due to the completely unsupervised nature of their training. Existing methods for gaining such steerability collect human labels of the relative quality of model generations and fine-tune the unsupervised LM to align with these preferences, often with reinforcement learning from human feedback (RLHF). However, RLHF is a complex and often unstable procedure, first fitting a reward model that reflects the human preferences, and then fine-tuning the large unsupervised LM using reinforcement learning to maximize this estimated reward without drifting too far from the original model. In this paper we introduce a new parameterization of the reward model in RLHF that enables extraction of the corresponding optimal policy in closed form, allowing us to solve the standard RLHF problem with only a simple classification loss. The resulting algorithm, which we call Direct Preference Optimization (DPO), is stable, performant, and computationally lightweight, eliminating the need for sampling from the LM during fine-tuning or performing significant hyperparameter tuning. Our experiments show that DPO can fine-tune LMs to align with human preferences as well as or better than existing methods. Notably, fine-tuning with DPO exceeds PPO-based RLHF in ability to control sentiment of generations, and matches or improves response quality in summarization and single-turn dialogue while being substantially simpler to implement and train.</p></blockquote> <p data-svelte-h="svelte-1s4mxs8">This post-training method was contributed by <a href="https://huggingface.co/kashif" rel="nofollow">Kashif Rasul</a> and later refactored by <a href="https://huggingface.co/qgallouedec" rel="nofollow">Quentin Gallouédec</a>.</p> <h2 class="relative group"><a id="quick-start" 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="#quick-start"><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>Quick start</span></h2> <p data-svelte-h="svelte-1cdam5">This example demonstrates how to train a language model using the <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> from TRL. We train a <a href="https://huggingface.co/Qwen/Qwen3-0.6B" rel="nofollow">Qwen 3 0.6B</a> model on the <a href="https://huggingface.co/datasets/openbmb/UltraFeedback" rel="nofollow">UltraFeedback dataset</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 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> trl <span class="hljs-keyword">import</span> DPOTrainer | |
| <span class="hljs-keyword">from</span> datasets <span class="hljs-keyword">import</span> load_dataset | |
| trainer = DPOTrainer( | |
| model=<span class="hljs-string">"Qwen/Qwen3-0.6B"</span>, | |
| train_dataset=load_dataset(<span class="hljs-string">"trl-lib/ultrafeedback_binarized"</span>, split=<span class="hljs-string">"train"</span>), | |
| ) | |
| trainer.train()<!-- HTML_TAG_END --></pre></div> <iframe src="https://trl-lib-trackio.hf.space/?project=trl-documentation&metrics=train*&sidebar=hidden&runs=dpo_qwen3-0.6B_ultrafeedback" style="width: 100%; min-width: 300px; max-width: 800px;" height="830" frameborder="0"></iframe> <h2 class="relative group"><a id="expected-dataset-type-and-format" 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="#expected-dataset-type-and-format"><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>Expected dataset type and format</span></h2> <p data-svelte-h="svelte-qszqkj">DPO requires a <a href="dataset_formats#preference">preference</a> dataset. The <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> is compatible with both <a href="dataset_formats#standard">standard</a> and <a href="dataset_formats#conversational">conversational</a> dataset formats. When provided with a conversational dataset, the trainer will automatically apply the chat template to the 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 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-comment"># Standard format</span> | |
| <span class="hljs-comment">## Explicit prompt (recommended)</span> | |
| preference_example = {<span class="hljs-string">"prompt"</span>: <span class="hljs-string">"The sky is"</span>, <span class="hljs-string">"chosen"</span>: <span class="hljs-string">" blue."</span>, <span class="hljs-string">"rejected"</span>: <span class="hljs-string">" green."</span>} | |
| <span class="hljs-comment"># Implicit prompt</span> | |
| preference_example = {<span class="hljs-string">"chosen"</span>: <span class="hljs-string">"The sky is blue."</span>, <span class="hljs-string">"rejected"</span>: <span class="hljs-string">"The sky is green."</span>} | |
| <span class="hljs-comment"># Conversational format</span> | |
| <span class="hljs-comment">## Explicit prompt (recommended)</span> | |
| preference_example = {<span class="hljs-string">"prompt"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"user"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"What color is the sky?"</span>}], | |
| <span class="hljs-string">"chosen"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"assistant"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"It is blue."</span>}], | |
| <span class="hljs-string">"rejected"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"assistant"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"It is green."</span>}]} | |
| <span class="hljs-comment">## Implicit prompt</span> | |
| preference_example = {<span class="hljs-string">"chosen"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"user"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"What color is the sky?"</span>}, | |
| {<span class="hljs-string">"role"</span>: <span class="hljs-string">"assistant"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"It is blue."</span>}], | |
| <span class="hljs-string">"rejected"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"user"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"What color is the sky?"</span>}, | |
| {<span class="hljs-string">"role"</span>: <span class="hljs-string">"assistant"</span>, <span class="hljs-string">"content"</span>: <span class="hljs-string">"It is green."</span>}]}<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-19drk8m">If your dataset is not in one of these formats, you can preprocess it to convert it into the expected format. Here is an example with the <a href="https://huggingface.co/datasets/Vezora/Code-Preference-Pairs" rel="nofollow">Vezora/Code-Preference-Pairs</a> 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 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> datasets <span class="hljs-keyword">import</span> load_dataset | |
| dataset = load_dataset(<span class="hljs-string">"Vezora/Code-Preference-Pairs"</span>) | |
| <span class="hljs-keyword">def</span> <span class="hljs-title function_">preprocess_function</span>(<span class="hljs-params">example</span>): | |
| <span class="hljs-keyword">return</span> { | |
| <span class="hljs-string">"prompt"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"user"</span>, <span class="hljs-string">"content"</span>: example[<span class="hljs-string">"input"</span>]}], | |
| <span class="hljs-string">"chosen"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"assistant"</span>, <span class="hljs-string">"content"</span>: example[<span class="hljs-string">"accepted"</span>]}], | |
| <span class="hljs-string">"rejected"</span>: [{<span class="hljs-string">"role"</span>: <span class="hljs-string">"assistant"</span>, <span class="hljs-string">"content"</span>: example[<span class="hljs-string">"rejected"</span>]}], | |
| } | |
| dataset = dataset.<span class="hljs-built_in">map</span>(preprocess_function, remove_columns=[<span class="hljs-string">"instruction"</span>, <span class="hljs-string">"input"</span>, <span class="hljs-string">"accepted"</span>, <span class="hljs-string">"ID"</span>]) | |
| <span class="hljs-built_in">print</span>(<span class="hljs-built_in">next</span>(<span class="hljs-built_in">iter</span>(dataset[<span class="hljs-string">"train"</span>])))<!-- HTML_TAG_END --></pre></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-punctuation">{</span> | |
| <span class="hljs-attr">"prompt"</span><span class="hljs-punctuation">:</span> <span class="hljs-punctuation">[</span><span class="hljs-punctuation">{</span><span class="hljs-attr">"role"</span><span class="hljs-punctuation">:</span> <span class="hljs-string">"user"</span><span class="hljs-punctuation">,</span> <span class="hljs-attr">"content"</span><span class="hljs-punctuation">:</span> <span class="hljs-string">"Create a nested loop to print every combination of numbers [...]"</span><span class="hljs-punctuation">}</span><span class="hljs-punctuation">]</span><span class="hljs-punctuation">,</span> | |
| <span class="hljs-attr">"chosen"</span><span class="hljs-punctuation">:</span> <span class="hljs-punctuation">[</span><span class="hljs-punctuation">{</span><span class="hljs-attr">"role"</span><span class="hljs-punctuation">:</span> <span class="hljs-string">"assistant"</span><span class="hljs-punctuation">,</span> <span class="hljs-attr">"content"</span><span class="hljs-punctuation">:</span> <span class="hljs-string">"Here is an example of a nested loop in Python [...]"</span><span class="hljs-punctuation">}</span><span class="hljs-punctuation">]</span><span class="hljs-punctuation">,</span> | |
| <span class="hljs-attr">"rejected"</span><span class="hljs-punctuation">:</span> <span class="hljs-punctuation">[</span><span class="hljs-punctuation">{</span><span class="hljs-attr">"role"</span><span class="hljs-punctuation">:</span> <span class="hljs-string">"assistant"</span><span class="hljs-punctuation">,</span> <span class="hljs-attr">"content"</span><span class="hljs-punctuation">:</span> <span class="hljs-string">"Here is an example of a nested loop in Python [...]"</span><span class="hljs-punctuation">}</span><span class="hljs-punctuation">]</span><span class="hljs-punctuation">,</span> | |
| <span class="hljs-punctuation">}</span><!-- HTML_TAG_END --></pre></div> <h2 class="relative group"><a id="looking-deeper-into-the-dpo-method" 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="#looking-deeper-into-the-dpo-method"><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>Looking deeper into the DPO method</span></h2> <p data-svelte-h="svelte-1l1hbfo">Direct Preference Optimization (DPO) is a training method designed to align a language model with preference data. Instead of supervised input–output pairs, the model is trained on pairs of completions to the same prompt, where one completion is preferred over the other. The objective directly optimizes the model to widen the margin between the log-likelihoods of preferred and dispreferred completions, relative to a reference model, without requiring an explicit reward model. In practice, this is typically achieved by suppressing the likelihood of dispreferred completions rather than by increasing the likelihood of preferred ones.</p> <p data-svelte-h="svelte-5spfdn">This section breaks down how DPO works in practice, covering the key steps: <strong>preprocessing</strong> and <strong>loss computation</strong>.</p> <h3 class="relative group"><a id="preprocessing-and-tokenization" 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="#preprocessing-and-tokenization"><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>Preprocessing and tokenization</span></h3> <p data-svelte-h="svelte-1lt8syj">During training, each example is expected to contain a prompt along with a preferred (<code>chosen</code>) and a dispreferred (<code>rejected</code>) completion. For more details on the expected formats, see <a href="dataset_formats">Dataset formats</a>. | |
| The <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> tokenizes each input using the model’s tokenizer.</p> <h3 class="relative group"><a id="computing-the-loss" 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="#computing-the-loss"><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>Computing the loss</span></h3> <p data-svelte-h="svelte-1en6j9l"><img src="https://huggingface.co/datasets/trl-lib/documentation-images/resolve/main/dpo_figure.png" alt="dpo_figure"></p> <p>The loss used in DPO is defined as follows:<!-- HTML_TAG_START --><span class="katex-display"><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML" display="block"><semantics><mrow><msub><mi mathvariant="script">L</mi><mrow><mi mathvariant="normal">D</mi><mi mathvariant="normal">P</mi><mi mathvariant="normal">O</mi></mrow></msub><mo stretchy="false">(</mo><mi>θ</mi><mo stretchy="false">)</mo><mo>=</mo><mo>−</mo><msub><mi mathvariant="double-struck">E</mi><mrow><mo stretchy="false">(</mo><mi>x</mi><mo separator="true">,</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">+</mo></msup><mo separator="true">,</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">−</mo></msup><mo stretchy="false">)</mo></mrow></msub><mtext> </mtext><mrow><mo fence="true">[</mo><mi>log</mi><mo></mo><mi>σ</mi><mtext> </mtext><mrow><mo fence="true">(</mo><mi>β</mi><mo fence="false" stretchy="true" minsize="1.8em" maxsize="1.8em">(</mo><mi>log</mi><mo></mo><mfrac><mrow><msub><mi>π</mi><mi>θ</mi></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">+</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow><mrow><msub><mi>π</mi><mrow><mi mathvariant="normal">r</mi><mi mathvariant="normal">e</mi><mi mathvariant="normal">f</mi></mrow></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">+</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow></mfrac><mo>−</mo><mi>log</mi><mo></mo><mfrac><mrow><msub><mi>π</mi><mi>θ</mi></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">−</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow><mrow><msub><mi>π</mi><mrow><mi mathvariant="normal">r</mi><mi mathvariant="normal">e</mi><mi mathvariant="normal">f</mi></mrow></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">−</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow></mfrac><mo fence="false" stretchy="true" minsize="1.8em" maxsize="1.8em">)</mo><mo fence="true">)</mo></mrow><mo fence="true">]</mo></mrow></mrow><annotation encoding="application/x-tex"> | |
| \mathcal{L}_{\mathrm{DPO}}(\theta) = -\mathbb{E}_{(x,y^{+},y^{-})}\!\left[\log \sigma\!\left(\beta\Big(\log\frac{\pi_{\theta}(y^{+}\!\mid x)}{\pi_{\mathrm{ref}}(y^{+}\!\mid x)}-\log \frac{\pi_{\theta}(y^{-}\!\mid x)}{\pi_{\mathrm{ref}}(y^{-}\!\mid x)}\Big)\right)\right] | |
| </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:1em;vertical-align:-0.25em;"></span><span class="mord"><span class="mord mathcal">L</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3283em;"><span style="top:-2.55em;margin-left:0em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathrm mtight">DPO</span></span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mopen">(</span><span class="mord mathnormal" style="margin-right:0.02778em;">θ</span><span class="mclose">)</span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mrel">=</span><span class="mspace" style="margin-right:0.2778em;"></span></span><span class="base"><span class="strut" style="height:2.4em;vertical-align:-0.95em;"></span><span class="mord">−</span><span class="mord"><span class="mord mathbb">E</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3448em;"><span style="top:-2.5198em;margin-left:0em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mopen mtight">(</span><span class="mord mathnormal mtight">x</span><span class="mpunct mtight">,</span><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7027em;"><span style="top:-2.786em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight">+</span></span></span></span></span></span></span></span></span><span class="mpunct mtight">,</span><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7027em;"><span style="top:-2.786em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight">−</span></span></span></span></span></span></span></span></span><span class="mclose mtight">)</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.3552em;"><span></span></span></span></span></span></span><span class="mspace" style="margin-right:-0.1667em;"></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="minner"><span class="mopen delimcenter" style="top:0em;"><span class="delimsizing size3">[</span></span><span class="mop">lo<span style="margin-right:0.01389em;">g</span></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mord mathnormal" style="margin-right:0.03588em;">σ</span><span class="mspace" style="margin-right:-0.1667em;"></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="minner"><span class="mopen delimcenter" style="top:0em;"><span class="delimsizing size3">(</span></span><span class="mord mathnormal" style="margin-right:0.05278em;">β</span><span class="mord"><span class="delimsizing size2">(</span></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mop">lo<span style="margin-right:0.01389em;">g</span></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mord"><span class="mopen nulldelimiter"></span><span class="mfrac"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:1.4483em;"><span style="top:-2.314em;"><span class="pstrut" style="height:3em;"></span><span class="mord"><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3361em;"><span style="top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathrm mtight" style="margin-right:0.07778em;">ref</span></span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mopen">(</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.6973em;"><span style="top:-2.989em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">+</span></span></span></span></span></span></span></span></span><span class="mspace" style="margin-right:-0.1667em;"></span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mrel">∣</span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mord mathnormal">x</span><span class="mclose">)</span></span></span><span style="top:-3.23em;"><span class="pstrut" style="height:3em;"></span><span class="frac-line" style="border-bottom-width:0.04em;"></span></span><span style="top:-3.677em;"><span class="pstrut" style="height:3em;"></span><span class="mord"><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3361em;"><span style="top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.02778em;">θ</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mopen">(</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7713em;"><span style="top:-3.063em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">+</span></span></span></span></span></span></span></span></span><span class="mspace" style="margin-right:-0.1667em;"></span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mrel">∣</span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mord mathnormal">x</span><span class="mclose">)</span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.936em;"><span></span></span></span></span></span><span class="mclose nulldelimiter"></span></span><span class="mspace" style="margin-right:0.2222em;"></span><span class="mbin">−</span><span class="mspace" style="margin-right:0.2222em;"></span><span class="mop">lo<span style="margin-right:0.01389em;">g</span></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mord"><span class="mopen nulldelimiter"></span><span class="mfrac"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:1.4483em;"><span style="top:-2.314em;"><span class="pstrut" style="height:3em;"></span><span class="mord"><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3361em;"><span style="top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathrm mtight" style="margin-right:0.07778em;">ref</span></span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mopen">(</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.6973em;"><span style="top:-2.989em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">−</span></span></span></span></span></span></span></span></span><span class="mspace" style="margin-right:-0.1667em;"></span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mrel">∣</span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mord mathnormal">x</span><span class="mclose">)</span></span></span><span style="top:-3.23em;"><span class="pstrut" style="height:3em;"></span><span class="frac-line" style="border-bottom-width:0.04em;"></span></span><span style="top:-3.677em;"><span class="pstrut" style="height:3em;"></span><span class="mord"><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3361em;"><span style="top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.02778em;">θ</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span><span class="mopen">(</span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7713em;"><span style="top:-3.063em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight">−</span></span></span></span></span></span></span></span></span><span class="mspace" style="margin-right:-0.1667em;"></span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mrel">∣</span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mord mathnormal">x</span><span class="mclose">)</span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.936em;"><span></span></span></span></span></span><span class="mclose nulldelimiter"></span></span><span class="mord"><span class="delimsizing size2">)</span></span><span class="mclose delimcenter" style="top:0em;"><span class="delimsizing size3">)</span></span></span><span class="mclose delimcenter" style="top:0em;"><span class="delimsizing size3">]</span></span></span></span></span></span></span><!-- HTML_TAG_END --></p> <p>where <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mi>x</mi></mrow><annotation encoding="application/x-tex"> x </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.4306em;"></span><span class="mord mathnormal">x</span></span></span></span><!-- HTML_TAG_END --> is the prompt, <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msup><mi>y</mi><mo>+</mo></msup></mrow><annotation encoding="application/x-tex"> y^+ </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.9658em;vertical-align:-0.1944em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7713em;"><span style="top:-3.063em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mbin mtight">+</span></span></span></span></span></span></span></span></span></span></span><!-- HTML_TAG_END --> is the preferred completion and <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msup><mi>y</mi><mo>−</mo></msup></mrow><annotation encoding="application/x-tex"> y^- </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.9658em;vertical-align:-0.1944em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7713em;"><span style="top:-3.063em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mbin mtight">−</span></span></span></span></span></span></span></span></span></span></span><!-- HTML_TAG_END --> is the dispreferred completion. <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msub><mi>π</mi><mi>θ</mi></msub></mrow><annotation encoding="application/x-tex"> \pi_{\theta} </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.5806em;vertical-align:-0.15em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3361em;"><span style="top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.02778em;">θ</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span></span></span></span><!-- HTML_TAG_END --> is the policy model being trained, <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><msub><mi>π</mi><mrow><mi mathvariant="normal">r</mi><mi mathvariant="normal">e</mi><mi mathvariant="normal">f</mi></mrow></msub></mrow><annotation encoding="application/x-tex"> \pi_{\mathrm{ref}} </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.5806em;vertical-align:-0.15em;"></span><span class="mord"><span class="mord mathnormal" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3361em;"><span style="top:-2.55em;margin-left:-0.0359em;margin-right:0.05em;"><span class="pstrut" style="height:2.7em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathrm mtight" style="margin-right:0.07778em;">ref</span></span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.15em;"><span></span></span></span></span></span></span></span></span></span><!-- HTML_TAG_END --> is the reference model, <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mi>σ</mi></mrow><annotation encoding="application/x-tex"> \sigma </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.4306em;"></span><span class="mord mathnormal" style="margin-right:0.03588em;">σ</span></span></span></span><!-- HTML_TAG_END --> is the sigmoid function, and <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mi>β</mi><mo>></mo><mn>0</mn></mrow><annotation encoding="application/x-tex"> \beta > 0 </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:0.8889em;vertical-align:-0.1944em;"></span><span class="mord mathnormal" style="margin-right:0.05278em;">β</span><span class="mspace" style="margin-right:0.2778em;"></span><span class="mrel">></span><span class="mspace" style="margin-right:0.2778em;"></span></span><span class="base"><span class="strut" style="height:0.6444em;"></span><span class="mord">0</span></span></span></span><!-- HTML_TAG_END --> is a hyperparameter that controls the strength of the preference signal.</p> <h4 class="relative group"><a id="loss-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="#loss-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>Loss Types</span></h4> <p data-svelte-h="svelte-14f6eh7">Several formulations of the objective have been proposed in the literature. Initially, the objective of DPO was defined as presented above.</p> <table data-svelte-h="svelte-2usb9i"><thead><tr><th><code>loss_type=</code></th> <th>Description</th></tr></thead> <tbody><tr><td><code>"sigmoid"</code> (default)</td> <td>Given the preference data, we can fit a binary classifier according to the Bradley-Terry model and in fact the <a href="https://huggingface.co/papers/2305.18290" rel="nofollow">DPO</a> authors propose the sigmoid loss on the normalized likelihood via the <code>logsigmoid</code> to fit a logistic regression.</td></tr> <tr><td><code>"hinge"</code></td> <td>The <a href="https://huggingface.co/papers/2309.06657" rel="nofollow">RSO</a> authors propose to use a hinge loss on the normalized likelihood from the <a href="https://huggingface.co/papers/2305.10425" rel="nofollow">SLiC</a> paper. In this case, the <code>beta</code> is the reciprocal of the margin.</td></tr> <tr><td><code>"ipo"</code></td> <td>The <a href="https://huggingface.co/papers/2310.12036" rel="nofollow">IPO</a> authors argue the logit transform can overfit and propose the identity transform to optimize preferences directly; TRL exposes this as <code>loss_type="ipo"</code>.</td></tr> <tr><td><code>"exo_pair"</code></td> <td>The <a href="https://huggingface.co/papers/2402.00856" rel="nofollow">EXO</a> authors propose reverse-KL preference optimization. <code>label_smoothing</code> must be strictly greater than <code>0.0</code>; a recommended value is <code>1e-3</code> (see Eq. 16 for the simplified pairwise variant). The full method uses <code>K>2</code> SFT completions and approaches PPO as <code>K</code> grows.</td></tr> <tr><td><code>"nca_pair"</code></td> <td>The <a href="https://huggingface.co/papers/2402.05369" rel="nofollow">NCA</a> authors shows that NCA optimizes the absolute likelihood for each response rather than the relative likelihood.</td></tr> <tr><td><code>"robust"</code></td> <td>The <a href="https://huggingface.co/papers/2403.00409" rel="nofollow">Robust DPO</a> authors propose an unbiased DPO loss under noisy preferences. Use <code>label_smoothing</code> in <a href="/docs/trl/pr_5607/en/dpo_trainer#trl.DPOConfig">DPOConfig</a> to model label-flip probability; valid values are in the range <code>[0.0, 0.5)</code>.</td></tr> <tr><td><code>"bco_pair"</code></td> <td>The <a href="https://huggingface.co/papers/2404.04656" rel="nofollow">BCO</a> authors train a binary classifier whose logit serves as a reward so that the classifier maps {prompt, chosen completion} pairs to 1 and {prompt, rejected completion} pairs to 0. For unpaired data, we recommend the dedicated <a href="/docs/trl/pr_5607/en/bco_trainer#trl.experimental.bco.BCOTrainer">experimental.bco.BCOTrainer</a>.</td></tr> <tr><td><code>"sppo_hard"</code></td> <td>The <a href="https://huggingface.co/papers/2405.00675" rel="nofollow">SPPO</a> authors claim that SPPO is capable of solving the Nash equilibrium iteratively by pushing the chosen rewards to be as large as 1/2 and the rejected rewards to be as small as -1/2 and can alleviate data sparsity issues. The implementation approximates this algorithm by employing hard label probabilities, assigning 1 to the winner and 0 to the loser.</td></tr> <tr><td><code>"aot"</code> or <code>loss_type="aot_unpaired"</code></td> <td>The <a href="https://huggingface.co/papers/2406.05882" rel="nofollow">AOT</a> authors propose Distributional Preference Alignment via Optimal Transport. <code>loss_type="aot"</code> is for paired data; <code>loss_type="aot_unpaired"</code> is for unpaired data. Both enforce stochastic dominance via sorted quantiles; larger per-GPU batch sizes help.</td></tr> <tr><td><code>"apo_zero"</code> or <code>loss_type="apo_down"</code></td> <td>The <a href="https://huggingface.co/papers/2408.06266" rel="nofollow">APO</a> method introduces an anchored objective. <code>apo_zero</code> boosts winners and downweights losers (useful when the model underperforms the winners). <code>apo_down</code> downweights both, with stronger pressure on losers (useful when the model already outperforms winners).</td></tr> <tr><td><code>"discopop"</code></td> <td>The <a href="https://huggingface.co/papers/2406.08414" rel="nofollow">DiscoPOP</a> paper uses LLMs to discover more efficient offline preference optimization losses. In the paper the proposed DiscoPOP loss (which is a log-ratio modulated loss) outperformed other optimization losses on different tasks (IMDb positive text generation, Reddit TLDR summarization, and Alpaca Eval 2.0).</td></tr> <tr><td><code>"sft"</code></td> <td>SFT (Supervised Fine-Tuning) loss is the negative log likelihood loss, used to train the model to generate preferred responses.</td></tr></tbody></table> <h2 class="relative group"><a id="logged-metrics" 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="#logged-metrics"><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>Logged metrics</span></h2> <p data-svelte-h="svelte-132s7j9">While training and evaluating we record the following reward metrics:</p> <ul><li data-svelte-h="svelte-jaaxpo"><code>global_step</code>: The total number of optimizer steps taken so far.</li> <li data-svelte-h="svelte-yo24c5"><code>epoch</code>: The current epoch number, based on dataset iteration.</li> <li data-svelte-h="svelte-16mlmr8"><code>num_tokens</code>: The total number of tokens processed so far.</li> <li data-svelte-h="svelte-7e9cei"><code>loss</code>: The average cross-entropy loss computed over non-masked tokens in the current logging interval.</li> <li data-svelte-h="svelte-1nxs1qk"><code>entropy</code>: The average entropy of the model’s predicted token distribution over non-masked tokens.</li> <li data-svelte-h="svelte-1818mwl"><code>mean_token_accuracy</code>: The proportion of non-masked tokens for which the model’s top-1 prediction matches the token from the chosen completion.</li> <li data-svelte-h="svelte-1rd9kbl"><code>learning_rate</code>: The current learning rate, which may change dynamically if a scheduler is used.</li> <li data-svelte-h="svelte-1jpw61c"><code>grad_norm</code>: The L2 norm of the gradients, computed before gradient clipping.</li> <li data-svelte-h="svelte-66fe9i"><code>logits/chosen</code>: The average logit values assigned by the model to the tokens in the chosen completion.</li> <li data-svelte-h="svelte-fxc0s6"><code>logits/rejected</code>: The average logit values assigned by the model to the tokens in the rejected completion.</li> <li data-svelte-h="svelte-1tnjo7i"><code>logps/chosen</code>: The average log-probability assigned by the model to the tokens in the chosen completion.</li> <li data-svelte-h="svelte-1s9v0ay"><code>logps/rejected</code>: The average log-probability assigned by the model to the tokens in the rejected completion.</li> <li><code data-svelte-h="svelte-l0e5jv">rewards/chosen</code>: The average implicit reward computed for the chosen completion, computed as <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mi>β</mi><mi>log</mi><mo></mo><mfrac><mrow><msub><mi>π</mi><mi>θ</mi></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">+</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow><mrow><msub><mi>π</mi><mrow><mi mathvariant="normal">r</mi><mi mathvariant="normal">e</mi><mi mathvariant="normal">f</mi></mrow></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">+</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow></mfrac></mrow><annotation encoding="application/x-tex"> \beta \log \frac{\pi_{\theta}(y^{+}\!\mid x)}{\pi_{\mathrm{ref}}(y^{+}\!\mid x)} </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:1.5984em;vertical-align:-0.52em;"></span><span class="mord mathnormal" style="margin-right:0.05278em;">β</span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mop">lo<span style="margin-right:0.01389em;">g</span></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mord"><span class="mopen nulldelimiter"></span><span class="mfrac"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:1.0784em;"><span style="top:-2.655em;"><span class="pstrut" style="height:3em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3448em;"><span style="top:-2.3488em;margin-left:-0.0359em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathrm mtight" style="margin-right:0.07778em;">ref</span></span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.1512em;"><span></span></span></span></span></span></span><span class="mopen mtight">(</span><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7027em;"><span style="top:-2.786em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight">+</span></span></span></span></span></span></span></span></span><span class="mspace mtight" style="margin-right:-0.1952em;"></span><span class="mrel mtight">∣</span><span class="mord mathnormal mtight">x</span><span class="mclose mtight">)</span></span></span></span><span style="top:-3.23em;"><span class="pstrut" style="height:3em;"></span><span class="frac-line" style="border-bottom-width:0.04em;"></span></span><span style="top:-3.485em;"><span class="pstrut" style="height:3em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3448em;"><span style="top:-2.3488em;margin-left:-0.0359em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.02778em;">θ</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.1512em;"><span></span></span></span></span></span></span><span class="mopen mtight">(</span><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.8477em;"><span style="top:-2.931em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight">+</span></span></span></span></span></span></span></span></span><span class="mspace mtight" style="margin-right:-0.1952em;"></span><span class="mrel mtight">∣</span><span class="mord mathnormal mtight">x</span><span class="mclose mtight">)</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.52em;"><span></span></span></span></span></span><span class="mclose nulldelimiter"></span></span></span></span></span><!-- HTML_TAG_END -->.</li> <li><code data-svelte-h="svelte-xqs1fr">rewards/rejected</code>: The average implicit reward computed for the rejected completion, computed as <!-- HTML_TAG_START --><span class="katex"><span class="katex-mathml"><math xmlns="http://www.w3.org/1998/Math/MathML"><semantics><mrow><mi>β</mi><mi>log</mi><mo></mo><mfrac><mrow><msub><mi>π</mi><mi>θ</mi></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">−</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow><mrow><msub><mi>π</mi><mrow><mi mathvariant="normal">r</mi><mi mathvariant="normal">e</mi><mi mathvariant="normal">f</mi></mrow></msub><mo stretchy="false">(</mo><msup><mi>y</mi><mo lspace="0em" rspace="0em">−</mo></msup><mtext> </mtext><mo>∣</mo><mi>x</mi><mo stretchy="false">)</mo></mrow></mfrac></mrow><annotation encoding="application/x-tex"> \beta \log \frac{\pi_{\theta}(y^{-}\!\mid x)}{\pi_{\mathrm{ref}}(y^{-}\!\mid x)} </annotation></semantics></math></span><span class="katex-html" aria-hidden="true"><span class="base"><span class="strut" style="height:1.5984em;vertical-align:-0.52em;"></span><span class="mord mathnormal" style="margin-right:0.05278em;">β</span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mop">lo<span style="margin-right:0.01389em;">g</span></span><span class="mspace" style="margin-right:0.1667em;"></span><span class="mord"><span class="mopen nulldelimiter"></span><span class="mfrac"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:1.0784em;"><span style="top:-2.655em;"><span class="pstrut" style="height:3em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3448em;"><span style="top:-2.3488em;margin-left:-0.0359em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathrm mtight" style="margin-right:0.07778em;">ref</span></span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.1512em;"><span></span></span></span></span></span></span><span class="mopen mtight">(</span><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.7027em;"><span style="top:-2.786em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight">−</span></span></span></span></span></span></span></span></span><span class="mspace mtight" style="margin-right:-0.1952em;"></span><span class="mrel mtight">∣</span><span class="mord mathnormal mtight">x</span><span class="mclose mtight">)</span></span></span></span><span style="top:-3.23em;"><span class="pstrut" style="height:3em;"></span><span class="frac-line" style="border-bottom-width:0.04em;"></span></span><span style="top:-3.485em;"><span class="pstrut" style="height:3em;"></span><span class="sizing reset-size6 size3 mtight"><span class="mord mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">π</span><span class="msupsub"><span class="vlist-t vlist-t2"><span class="vlist-r"><span class="vlist" style="height:0.3448em;"><span style="top:-2.3488em;margin-left:-0.0359em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.02778em;">θ</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.1512em;"><span></span></span></span></span></span></span><span class="mopen mtight">(</span><span class="mord mtight"><span class="mord mathnormal mtight" style="margin-right:0.03588em;">y</span><span class="msupsub"><span class="vlist-t"><span class="vlist-r"><span class="vlist" style="height:0.8477em;"><span style="top:-2.931em;margin-right:0.0714em;"><span class="pstrut" style="height:2.5em;"></span><span class="sizing reset-size3 size1 mtight"><span class="mord mtight"><span class="mord mtight">−</span></span></span></span></span></span></span></span></span><span class="mspace mtight" style="margin-right:-0.1952em;"></span><span class="mrel mtight">∣</span><span class="mord mathnormal mtight">x</span><span class="mclose mtight">)</span></span></span></span></span><span class="vlist-s"></span></span><span class="vlist-r"><span class="vlist" style="height:0.52em;"><span></span></span></span></span></span><span class="mclose nulldelimiter"></span></span></span></span></span><!-- HTML_TAG_END -->.</li> <li data-svelte-h="svelte-5fr60y"><code>rewards/margins</code>: The average implicit reward margin between the chosen and rejected completions.</li> <li data-svelte-h="svelte-1c9sx89"><code>rewards/accuracies</code>: The proportion of examples where the implicit reward for the chosen completion is higher than that for the rejected completion.</li></ul> <h2 class="relative group"><a id="customization" 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="#customization"><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>Customization</span></h2> <h3 class="relative group"><a id="compatibility-and-constraints" 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="#compatibility-and-constraints"><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>Compatibility and constraints</span></h3> <p data-svelte-h="svelte-1mje9pc">Some argument combinations are intentionally restricted in the current <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> implementation:</p> <ul data-svelte-h="svelte-1uflo1q"><li><code>use_weighting=True</code> is not supported with <code>loss_type="aot"</code> or <code>loss_type="aot_unpaired"</code>.</li> <li>With <code>use_liger_kernel=True</code>: | |
| <ul><li>only a single <code>loss_type</code> is supported,</li> <li><code>compute_metrics</code> is not supported,</li> <li><code>precompute_ref_log_probs=True</code> is not supported.</li></ul></li> <li><code>sync_ref_model=True</code> is not supported when training with PEFT models that do not keep a standalone <code>ref_model</code>.</li> <li><code>sync_ref_model=True</code> cannot be combined with <code>precompute_ref_log_probs=True</code>.</li> <li><code>precompute_ref_log_probs=True</code> is not supported with <code>IterableDataset</code> (train or eval).</li></ul> <h3 class="relative group"><a id="multi-loss-combinations" 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="#multi-loss-combinations"><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>Multi-loss combinations</span></h3> <p data-svelte-h="svelte-5n3tdq">The DPO trainer supports combining multiple loss functions with different weights, enabling more sophisticated optimization strategies. This is particularly useful for implementing algorithms like MPO (Mixed Preference Optimization). MPO is a training approach that combines multiple optimization objectives, as described in the paper <a href="https://huggingface.co/papers/2411.10442" rel="nofollow">Enhancing the Reasoning Ability of Multimodal Large Language Models via Mixed Preference Optimization</a>.</p> <p data-svelte-h="svelte-1i4tgdd">To combine multiple losses, specify the loss types and corresponding weights as lists:</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-comment"># MPO: Combines DPO (sigmoid) for preference and BCO (bco_pair) for quality</span> | |
| training_args = DPOConfig( | |
| loss_type=[<span class="hljs-string">"sigmoid"</span>, <span class="hljs-string">"bco_pair"</span>, <span class="hljs-string">"sft"</span>], <span class="hljs-comment"># loss types to combine</span> | |
| loss_weights=[<span class="hljs-number">0.8</span>, <span class="hljs-number">0.2</span>, <span class="hljs-number">1.0</span>] <span class="hljs-comment"># corresponding weights, as used in the MPO paper</span> | |
| )<!-- HTML_TAG_END --></pre></div> <h3 class="relative group"><a id="model-initialization" 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="#model-initialization"><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>Model initialization</span></h3> <p data-svelte-h="svelte-c9trke">You can directly pass the kwargs of the <code>from_pretrained()</code> method to the <a href="/docs/trl/pr_5607/en/dpo_trainer#trl.DPOConfig">DPOConfig</a>. For example, if you want to load a model in a different precision, analogous to</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 -->model = AutoModelForCausalLM.from_pretrained(<span class="hljs-string">"Qwen/Qwen3-0.6B"</span>, dtype=torch.bfloat16)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1k4uevu">you can do so by passing the <code>model_init_kwargs={"dtype": torch.bfloat16}</code> argument to the <a href="/docs/trl/pr_5607/en/dpo_trainer#trl.DPOConfig">DPOConfig</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 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> trl <span class="hljs-keyword">import</span> DPOConfig | |
| training_args = DPOConfig( | |
| model_init_kwargs={<span class="hljs-string">"dtype"</span>: torch.bfloat16}, | |
| )<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-1su7aih">Note that all keyword arguments of <code>from_pretrained()</code> are supported.</p> <h3 class="relative group"><a id="train-adapters-with-peft" 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-adapters-with-peft"><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>Train adapters with PEFT</span></h3> <p data-svelte-h="svelte-t2zuq8">We support tight integration with 🤗 PEFT library, allowing any user to conveniently train adapters and share them on the Hub, rather than training the entire model.</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> datasets <span class="hljs-keyword">import</span> load_dataset | |
| <span class="hljs-keyword">from</span> trl <span class="hljs-keyword">import</span> DPOTrainer | |
| <span class="hljs-keyword">from</span> peft <span class="hljs-keyword">import</span> LoraConfig | |
| dataset = load_dataset(<span class="hljs-string">"trl-lib/ultrafeedback_binarized"</span>, split=<span class="hljs-string">"train"</span>) | |
| trainer = DPOTrainer( | |
| <span class="hljs-string">"Qwen/Qwen3-0.6B"</span>, | |
| train_dataset=dataset, | |
| peft_config=LoraConfig(), | |
| ) | |
| trainer.train()<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-m3geut">You can also continue training your <a href="https://huggingface.co/docs/peft/main/en/package_reference/peft_model#peft.PeftModel" rel="nofollow">PeftModel</a>. For that, first load a <code>PeftModel</code> outside <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> and pass it directly to the trainer without the <code>peft_config</code> argument being passed.</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> datasets <span class="hljs-keyword">import</span> load_dataset | |
| <span class="hljs-keyword">from</span> trl <span class="hljs-keyword">import</span> DPOTrainer | |
| <span class="hljs-keyword">from</span> peft <span class="hljs-keyword">import</span> AutoPeftModelForCausalLM | |
| model = AutoPeftModelForCausalLM.from_pretrained(<span class="hljs-string">"trl-lib/Qwen3-4B-LoRA"</span>, is_trainable=<span class="hljs-literal">True</span>) | |
| dataset = load_dataset(<span class="hljs-string">"trl-lib/ultrafeedback_binarized"</span>, split=<span class="hljs-string">"train"</span>) | |
| trainer = DPOTrainer( | |
| model=model, | |
| train_dataset=dataset, | |
| ) | |
| trainer.train()<!-- HTML_TAG_END --></pre></div> <blockquote class="tip"><p data-svelte-h="svelte-19zz1g2">When training adapters, you typically use a higher learning rate (≈1e‑5) than full fine-tuning since only new parameters are being learned.</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 -->DPOConfig(learning_rate=<span class="hljs-number">1e-5</span>, ...)<!-- HTML_TAG_END --></pre></div></blockquote> <h3 class="relative group"><a id="train-with-liger-kernel" 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-with-liger-kernel"><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>Train with Liger Kernel</span></h3> <p data-svelte-h="svelte-1i5ejat">Liger Kernel is a collection of Triton kernels for LLM training that boosts multi-GPU throughput by 20%, cuts memory use by 60% (enabling up to 4× longer context), and works seamlessly with tools like FlashAttention, PyTorch FSDP, and DeepSpeed. For more information, see <a href="liger_kernel_integration">Liger Kernel Integration</a>.</p> <h3 class="relative group"><a id="rapid-experimentation-for-dpo" 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="#rapid-experimentation-for-dpo"><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>Rapid Experimentation for DPO</span></h3> <p data-svelte-h="svelte-uji4yk">RapidFire AI is an open-source experimentation engine that sits on top of TRL and lets you launch multiple DPO configurations at once, even on a single GPU. Instead of trying configurations sequentially, RapidFire lets you <strong>see all their learning curves earlier, stop underperforming runs, and clone promising ones with new settings in flight</strong> without restarting. For more information, see <a href="rapidfire_integration">RapidFire AI Integration</a>.</p> <h3 class="relative group"><a id="train-with-unsloth" 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-with-unsloth"><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>Train with Unsloth</span></h3> <p data-svelte-h="svelte-15c0opu">Unsloth is an open‑source framework for fine‑tuning and reinforcement learning that trains LLMs (like Llama, Mistral, Gemma, DeepSeek, and more) up to 2× faster with up to 70% less VRAM, while providing a streamlined, Hugging Face–compatible workflow for training, evaluation, and deployment. For more information, see <a href="unsloth_integration">Unsloth Integration</a>.</p> <h2 class="relative group"><a id="tool-calling-with-dpo" 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="#tool-calling-with-dpo"><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>Tool Calling with DPO</span></h2> <p data-svelte-h="svelte-qnusbg">The <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> fully supports fine-tuning models with <em>tool calling</em> capabilities. In this case, each dataset example should include:</p> <ul data-svelte-h="svelte-avzfup"><li>The conversation messages (prompt, chosen and rejected), including any tool calls (<code>tool_calls</code>) and tool responses (<code>tool</code> role messages)</li> <li>The list of available tools in the <code>tools</code> column, typically provided as JSON schemas</li></ul> <p data-svelte-h="svelte-vl4ede">For details on the expected dataset structure, see the <a href="dataset_formats#tool-calling">Dataset Format — Tool Calling</a> section.</p> <h2 class="relative group"><a id="training-vision-language-models" 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="#training-vision-language-models"><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>Training Vision Language Models</span></h2> <p data-svelte-h="svelte-ubg1n6"><a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> fully supports training Vision-Language Models (VLMs). To train a VLM, provide a dataset with either an <code>image</code> column (single image per sample) or an <code>images</code> column (list of images per sample). For more information on the expected dataset structure, see the <a href="dataset_formats#vision-dataset">Dataset Format — Vision Dataset</a> section. | |
| An example of such a dataset is the <a href="https://huggingface.co/datasets/HuggingFaceH4/rlaif-v_formatted" rel="nofollow">RLAIF-V Dataset</a> 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 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> trl <span class="hljs-keyword">import</span> DPOConfig, DPOTrainer | |
| <span class="hljs-keyword">from</span> datasets <span class="hljs-keyword">import</span> load_dataset | |
| trainer = DPOTrainer( | |
| model=<span class="hljs-string">"Qwen/Qwen2.5-VL-3B-Instruct"</span>, | |
| args=DPOConfig(max_length=<span class="hljs-literal">None</span>), | |
| train_dataset=load_dataset(<span class="hljs-string">"HuggingFaceH4/rlaif-v_formatted"</span>, split=<span class="hljs-string">"train"</span>), | |
| ) | |
| trainer.train()<!-- HTML_TAG_END --></pre></div> <blockquote class="tip"><p data-svelte-h="svelte-f7wiem">For VLMs, truncating may remove image tokens, leading to errors during training. To avoid this, set <code>max_length=None</code> in the <a href="/docs/trl/pr_5607/en/dpo_trainer#trl.DPOConfig">DPOConfig</a>. This allows the model to process the full sequence length without truncating image 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 -->DPOConfig(max_length=<span class="hljs-literal">None</span>, ...)<!-- HTML_TAG_END --></pre></div> <p data-svelte-h="svelte-7mhisv">Only use <code>max_length</code> when you’ve verified that truncation won’t remove image tokens for the entire dataset.</p></blockquote> <h2 class="relative group"><a id="trl.DPOTrainer" 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="#trl.DPOTrainer"><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>DPOTrainer</span></h2> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"> <div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="trl.DPOTrainer"><!-- HTML_TAG_START --><h3 class="!m-0"><span class="flex-1 break-all md:text-lg bg-gradient-to-r px-2.5 py-1.5 rounded-xl from-indigo-50/70 to-white dark:from-gray-900 dark:to-gray-950 dark:text-indigo-300 text-indigo-700"><svg class="mr-1.5 text-indigo-500 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" focusable="false" role="img" width=".8em" height=".8em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 24 24"><path class="uim-quaternary" d="M20.23 7.24L12 12L3.77 7.24a1.98 1.98 0 0 1 .7-.71L11 2.76c.62-.35 1.38-.35 2 0l6.53 3.77c.29.173.531.418.7.71z" opacity=".25" fill="currentColor"></path><path class="uim-tertiary" d="M12 12v9.5a2.09 2.09 0 0 1-.91-.21L4.5 17.48a2.003 2.003 0 0 1-1-1.73v-7.5a2.06 2.06 0 0 1 .27-1.01L12 12z" opacity=".5" fill="currentColor"></path><path class="uim-primary" d="M20.5 8.25v7.5a2.003 2.003 0 0 1-1 1.73l-6.62 3.82c-.275.13-.576.198-.88.2V12l8.23-4.76c.175.308.268.656.27 1.01z" fill="currentColor"></path></svg><span class="font-light">class</span> <span class="font-medium">trl.</span><span class="font-semibold">DPOTrainer</span></span></h3><!-- HTML_TAG_END --> <a id="trl.DPOTrainer" class="header-link invisible with-hover:group-hover:visible pr-2" href="#trl.DPOTrainer"><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></a> <a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/trl/blob/vr_5607/trl/trainer/dpo_trainer.py#L406" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span class="hidden md:block mx-0.5 hover:!underline" data-svelte-h="svelte-122apf4">source</span> <span data-svelte-h="svelte-x0xyl0">></span></a></span> <p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span data-svelte-h="svelte-8mvn6a">(</span> <span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">model<span class="opacity-60">: str | PreTrainedModel | PeftModel</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ref_model<span class="opacity-60">: transformers.modeling_utils.PreTrainedModel | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">args<span class="opacity-60">: trl.trainer.dpo_config.DPOConfig | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">data_collator<span class="opacity-60">: collections.abc.Callable[[list[typing.Any]], dict[str, typing.Any]] | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">train_dataset<span class="opacity-60">: datasets.arrow_dataset.Dataset | datasets.iterable_dataset.IterableDataset | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_dataset<span class="opacity-60">: datasets.arrow_dataset.Dataset | datasets.iterable_dataset.IterableDataset | dict[str, datasets.arrow_dataset.Dataset | datasets.iterable_dataset.IterableDataset] | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">processing_class<span class="opacity-60">: transformers.tokenization_utils_base.PreTrainedTokenizerBase | transformers.processing_utils.ProcessorMixin | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">compute_metrics<span class="opacity-60">: collections.abc.Callable[[transformers.trainer_utils.EvalPrediction], dict] | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">callbacks<span class="opacity-60">: list[transformers.trainer_callback.TrainerCallback] | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">optimizers<span class="opacity-60">: tuple = (None, None)</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">peft_config<span class="opacity-60">: PeftConfig | None = None</span></span> </span> <span data-svelte-h="svelte-1jq0pl7">)</span> </p> <div class="!mb-10 relative docstring-details "> <p class="flex items-center font-semibold !mt-2 !mb-2 text-gray-800" data-svelte-h="svelte-lt6pb6">Parameters <span class="flex-auto border-t-2 border-gray-100 dark:border-gray-700 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.model" class="header-link block pr-0.5 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="#trl.DPOTrainer.model"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>model</strong> (<code>str</code> or <a href="https://huggingface.co/docs/transformers/main/en/main_classes/model#transformers.PreTrainedModel" rel="nofollow">PreTrainedModel</a> or <a href="https://huggingface.co/docs/peft/main/en/package_reference/peft_model#peft.PeftModel" rel="nofollow">PeftModel</a>) — | |
| Model to be trained. Can be either:</p> | |
| <ul> | |
| <li>A string, being the <em>model id</em> of a pretrained model hosted inside a model repo on huggingface.co, or a | |
| path to a <em>directory</em> containing model weights saved using | |
| <a href="https://huggingface.co/docs/transformers/main/en/main_classes/model#transformers.PreTrainedModel.save_pretrained" rel="nofollow">save_pretrained</a>, e.g., <code>'./my_model_directory/'</code>. The model is loaded | |
| using <code><ModelArchitecture>.from_pretrained</code> (where <code><ModelArchitecture></code> is derived from the model | |
| config) with the keyword arguments in <code>args.model_init_kwargs</code>.</li> | |
| <li>A <a href="https://huggingface.co/docs/transformers/main/en/main_classes/model#transformers.PreTrainedModel" rel="nofollow">PreTrainedModel</a> object. Only causal language models are supported.</li> | |
| <li>A <a href="https://huggingface.co/docs/peft/main/en/package_reference/peft_model#peft.PeftModel" rel="nofollow">PeftModel</a> object. Only causal language models are supported.</li> | |
| </ul><!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.ref_model" class="header-link block pr-0.5 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="#trl.DPOTrainer.ref_model"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>ref_model</strong> (<a href="https://huggingface.co/docs/transformers/main/en/main_classes/model#transformers.PreTrainedModel" rel="nofollow">PreTrainedModel</a>, <em>optional</em>) — | |
| Reference model used to compute the reference log probabilities.</p> | |
| <ul> | |
| <li>If provided, this model is used directly as the reference policy.</li> | |
| <li>If <code>None</code>, the trainer will automatically use the initial policy corresponding to <code>model</code>, i.e. the model | |
| state before DPO training starts.</li> | |
| </ul><!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.args" class="header-link block pr-0.5 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="#trl.DPOTrainer.args"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>args</strong> (<a href="/docs/trl/pr_5607/en/dpo_trainer#trl.DPOConfig">DPOConfig</a>, <em>optional</em>) — | |
| Configuration for this trainer. If <code>None</code>, a default configuration is used.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.data_collator" class="header-link block pr-0.5 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="#trl.DPOTrainer.data_collator"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>data_collator</strong> (<code>DataCollator</code>, <em>optional</em>) — | |
| Function to use to form a batch from a list of elements of the processed <code>train_dataset</code> or <code>eval_dataset</code>. | |
| Will default to <code>DataCollatorForPreference</code> if the model is a language model and | |
| <code>DataCollatorForVisionPreference</code> if the model is a vision-language model. Custom | |
| collators must truncate sequences before padding; the trainer does not apply post-collation truncation.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.train_dataset" class="header-link block pr-0.5 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="#trl.DPOTrainer.train_dataset"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>train_dataset</strong> (<a href="https://huggingface.co/docs/datasets/main/en/package_reference/main_classes#datasets.Dataset" rel="nofollow">Dataset</a> or <a href="https://huggingface.co/docs/datasets/main/en/package_reference/main_classes#datasets.IterableDataset" rel="nofollow">IterableDataset</a>) — | |
| Dataset to use for training. This trainer supports both <a href="#language-modeling">language modeling</a> type and | |
| <a href="#prompt-completion">prompt-completion</a> type. The format of the samples can be either:</p> | |
| <ul> | |
| <li><a href="dataset_formats#standard">Standard</a>: Each sample contains plain text.</li> | |
| <li><a href="dataset_formats#conversational">Conversational</a>: Each sample contains structured messages (e.g., role | |
| and content).</li> | |
| </ul><!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.eval_dataset" class="header-link block pr-0.5 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="#trl.DPOTrainer.eval_dataset"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>eval_dataset</strong> (<a href="https://huggingface.co/docs/datasets/main/en/package_reference/main_classes#datasets.Dataset" rel="nofollow">Dataset</a>, <a href="https://huggingface.co/docs/datasets/main/en/package_reference/main_classes#datasets.IterableDataset" rel="nofollow">IterableDataset</a> or <code>dict[str, Dataset | IterableDataset]</code>) — | |
| Dataset to use for evaluation. It must meet the same requirements as <code>train_dataset</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.processing_class" class="header-link block pr-0.5 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="#trl.DPOTrainer.processing_class"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>processing_class</strong> (<a href="https://huggingface.co/docs/transformers/main/en/internal/tokenization_utils#transformers.PreTrainedTokenizerBase" rel="nofollow">PreTrainedTokenizerBase</a> or <a href="https://huggingface.co/docs/transformers/main/en/main_classes/processors#transformers.ProcessorMixin" rel="nofollow">ProcessorMixin</a>, <em>optional</em>) — | |
| Processing class used to process the data. The padding side must be set to “left”. If <code>None</code>, the | |
| processing class is loaded from the model’s name with <a href="https://huggingface.co/docs/transformers/main/en/model_doc/auto#transformers.AutoProcessor.from_pretrained" rel="nofollow">from_pretrained</a>. A | |
| padding token, <code>tokenizer.pad_token</code>, must be set. If the processing class has not set a padding token, | |
| <code>tokenizer.eos_token</code> will be used as the default.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.compute_metrics" class="header-link block pr-0.5 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="#trl.DPOTrainer.compute_metrics"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>compute_metrics</strong> (<code>Callable[[EvalPrediction], dict]</code>, <em>optional</em>) — | |
| The function that will be used to compute metrics at evaluation. Must take a | |
| <a href="https://huggingface.co/docs/transformers/main/en/internal/trainer_utils#transformers.EvalPrediction" rel="nofollow">EvalPrediction</a> and return a dictionary string to metric values. When passing | |
| <a href="/docs/trl/pr_5607/en/sft_trainer#trl.SFTConfig">SFTConfig</a> with <code>batch_eval_metrics</code> set to <code>True</code>, your <code>compute_metrics</code> function must take a boolean | |
| <code>compute_result</code> argument. This will be triggered after the last eval batch to signal that the function | |
| needs to calculate and return the global summary statistics rather than accumulating the batch-level | |
| statistics.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.callbacks" class="header-link block pr-0.5 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="#trl.DPOTrainer.callbacks"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>callbacks</strong> (list of <a href="https://huggingface.co/docs/transformers/main/en/main_classes/callback#transformers.TrainerCallback" rel="nofollow">TrainerCallback</a>, <em>optional</em>) — | |
| List of callbacks to customize the training loop. Will add those to the list of default callbacks detailed | |
| in <a href="https://huggingface.co/docs/transformers/main_classes/callback" rel="nofollow">here</a>.</p> | |
| <p>If you want to remove one of the default callbacks used, use the <a href="https://huggingface.co/docs/transformers/main/en/main_classes/trainer#transformers.Trainer.remove_callback" rel="nofollow">remove_callback</a> | |
| method.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.optimizers" class="header-link block pr-0.5 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="#trl.DPOTrainer.optimizers"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>optimizers</strong> (<code>tuple[torch.optim.Optimizer | None, torch.optim.lr_scheduler.LambdaLR | None]</code>, <em>optional</em>, defaults to <code>(None, None)</code>) — | |
| A tuple containing the optimizer and the scheduler to use. Will default to an instance of <code>AdamW</code> on your | |
| model and a scheduler given by <a href="https://huggingface.co/docs/transformers/main/en/main_classes/optimizer_schedules#transformers.get_linear_schedule_with_warmup" rel="nofollow">get_linear_schedule_with_warmup</a> controlled by <code>args</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.peft_config" class="header-link block pr-0.5 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="#trl.DPOTrainer.peft_config"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>peft_config</strong> (<a href="https://huggingface.co/docs/peft/main/en/package_reference/config#peft.PeftConfig" rel="nofollow">PeftConfig</a>, <em>optional</em>) — | |
| PEFT configuration used to wrap the model. If <code>None</code>, the model is not wrapped.<!-- HTML_TAG_END --> </span></span> </li></ul> </div></div> <p data-svelte-h="svelte-10hj3j9">Trainer for Direct Preference Optimization (DPO) method. This algorithm was initially proposed in the paper <a href="https://huggingface.co/papers/2305.18290" rel="nofollow">Direct | |
| Preference Optimization: Your Language Model is Secretly a Reward Model</a>. | |
| This class is a wrapper around the <a href="https://huggingface.co/docs/transformers/main/en/main_classes/trainer#transformers.Trainer" rel="nofollow">Trainer</a> class and inherits all of its attributes and methods.</p> <div class="relative group rounded-md"><a id="trl.DPOTrainer.example" class="header-link block pr-0.5 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="#trl.DPOTrainer.example"><span><svg class="text-smd" 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> <p data-svelte-h="svelte-11lpom8">Example:</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> trl <span class="hljs-keyword">import</span> DPOTrainer | |
| <span class="hljs-keyword">from</span> datasets <span class="hljs-keyword">import</span> load_dataset | |
| dataset = load_dataset(<span class="hljs-string">"trl-lib/ultrafeedback_binarized"</span>, split=<span class="hljs-string">"train"</span>) | |
| trainer = DPOTrainer( | |
| model=<span class="hljs-string">"Qwen/Qwen2.5-0.5B-Instruct"</span>, | |
| train_dataset=dataset, | |
| ) | |
| trainer.train()<!-- HTML_TAG_END --></pre></div></div> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"> <div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="trl.DPOTrainer.train"><!-- HTML_TAG_START --><h4 class="!m-0"><span class="flex-1 rounded-xl py-0.5 break-all bg-gradient-to-r from-blue-50/60 to-white dark:from-gray-900 dark:to-gray-950 text-blue-700 dark:text-blue-300 font-medium px-2"><svg width="1em" height="1em" viewBox="0 0 32 33" class="mr-1 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg"><path d="M5.80566 18.3545C4.90766 17.4565 4.90766 16.0005 5.80566 15.1025L14.3768 6.53142C15.2748 5.63342 16.7307 5.63342 17.6287 6.53142L26.1999 15.1025C27.0979 16.0005 27.0979 17.4565 26.1999 18.3545L17.6287 26.9256C16.7307 27.8236 15.2748 27.8236 14.3768 26.9256L5.80566 18.3545Z" fill="currentColor" fill-opacity="0.25"/><path fill-rule="evenodd" clip-rule="evenodd" d="M16.4801 13.9619C16.4801 12.9761 16.7467 12.5436 16.9443 12.3296C17.1764 12.078 17.5731 11.8517 18.2275 11.707C18.8821 11.5623 19.638 11.5342 20.4038 11.5582C20.7804 11.57 21.1341 11.5932 21.4719 11.6156L21.5263 11.6193C21.8195 11.6389 22.1626 11.6618 22.4429 11.6618V7.40825C22.3209 7.40825 22.1219 7.39596 21.7544 7.37149C21.4202 7.34925 20.9976 7.32115 20.5371 7.30672C19.6286 7.27824 18.4672 7.29779 17.3093 7.55377C16.1512 7.8098 14.8404 8.33724 13.8181 9.4452C12.7612 10.5907 12.2266 12.1236 12.2266 13.9619V15.0127H10.6836V19.2662H12.2266V26.6332H16.4801V19.2662H20.3394V15.0127H16.4801V13.9619Z" fill="currentColor"/></svg>train</span></h4><!-- HTML_TAG_END --> <a id="trl.DPOTrainer.train" class="header-link invisible with-hover:group-hover:visible pr-2" href="#trl.DPOTrainer.train"><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></a> <a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/trl/blob/vr_5607/transformers/trainer.py#L1323" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span class="hidden md:block mx-0.5 hover:!underline" data-svelte-h="svelte-122apf4">source</span> <span data-svelte-h="svelte-x0xyl0">></span></a></span> <p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span data-svelte-h="svelte-8mvn6a">(</span> <span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">resume_from_checkpoint<span class="opacity-60">: str | bool | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">trial<span class="opacity-60">: optuna.Trial | dict[str, Any] | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ignore_keys_for_eval<span class="opacity-60">: list[str] | None = None</span></span> </span> <span data-svelte-h="svelte-1jq0pl7">)</span> <span class="font-bold" data-svelte-h="svelte-1j6k10o">→</span> <span class="rounded hover:bg-gray-400 cursor-pointer"><!-- HTML_TAG_START --><script context="module">export const metadata = 'undefined';</script><span><code>~trainer_utils.TrainOutput</code></span><!-- HTML_TAG_END --></span></p> <div class="!mb-10 relative docstring-details "> <p class="flex items-center font-semibold !mt-2 !mb-2 text-gray-800" data-svelte-h="svelte-lt6pb6">Parameters <span class="flex-auto border-t-2 border-gray-100 dark:border-gray-700 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.train.resume_from_checkpoint" class="header-link block pr-0.5 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="#trl.DPOTrainer.train.resume_from_checkpoint"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>resume_from_checkpoint</strong> (<code>str</code> or <code>bool</code>, <em>optional</em>) — | |
| If a <code>str</code>, local path to a saved checkpoint as saved by a previous instance of <code>Trainer</code>. If a | |
| <code>bool</code> and equals <code>True</code>, load the last checkpoint in <em>args.output_dir</em> as saved by a previous instance | |
| of <code>Trainer</code>. If present, training will resume from the model/optimizer/scheduler states loaded here.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.train.trial" class="header-link block pr-0.5 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="#trl.DPOTrainer.train.trial"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>trial</strong> (<code>optuna.Trial</code> or <code>dict[str, Any]</code>, <em>optional</em>) — | |
| The trial run or the hyperparameter dictionary for hyperparameter search.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.train.ignore_keys_for_eval" class="header-link block pr-0.5 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="#trl.DPOTrainer.train.ignore_keys_for_eval"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>ignore_keys_for_eval</strong> (<code>list[str]</code>, <em>optional</em>) — | |
| A list of keys in the output of your model (if it is a dictionary) that should be ignored when | |
| gathering predictions for evaluation during the training.<!-- HTML_TAG_END --> </span></span> </li></ul> <div id="trl.DPOTrainer.train.returns" class="flex items-center font-semibold space-x-3 text-base !mt-0 !mb-0 text-gray-800 rounded "><p class="text-base">Returns</p> <!-- HTML_TAG_START --><script context="module">export const metadata = 'undefined';</script> | |
| <p><code>~trainer_utils.TrainOutput</code></p> | |
| <!-- HTML_TAG_END --> <span class="flex-auto border-t-2 border-gray-100 dark:border-gray-700"></span></div> <p class="text-base"><!-- HTML_TAG_START --><script context="module">export const metadata = 'undefined';</script> | |
| <p>Object containing the global step count, training loss, and metrics.</p> | |
| <!-- HTML_TAG_END --></p> </div></div> <p data-svelte-h="svelte-1cilnet">Main training entry point.</p></div> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"> <div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="trl.DPOTrainer.save_model"><!-- HTML_TAG_START --><h4 class="!m-0"><span class="flex-1 rounded-xl py-0.5 break-all bg-gradient-to-r from-blue-50/60 to-white dark:from-gray-900 dark:to-gray-950 text-blue-700 dark:text-blue-300 font-medium px-2"><svg width="1em" height="1em" viewBox="0 0 32 33" class="mr-1 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg"><path d="M5.80566 18.3545C4.90766 17.4565 4.90766 16.0005 5.80566 15.1025L14.3768 6.53142C15.2748 5.63342 16.7307 5.63342 17.6287 6.53142L26.1999 15.1025C27.0979 16.0005 27.0979 17.4565 26.1999 18.3545L17.6287 26.9256C16.7307 27.8236 15.2748 27.8236 14.3768 26.9256L5.80566 18.3545Z" fill="currentColor" fill-opacity="0.25"/><path fill-rule="evenodd" clip-rule="evenodd" d="M16.4801 13.9619C16.4801 12.9761 16.7467 12.5436 16.9443 12.3296C17.1764 12.078 17.5731 11.8517 18.2275 11.707C18.8821 11.5623 19.638 11.5342 20.4038 11.5582C20.7804 11.57 21.1341 11.5932 21.4719 11.6156L21.5263 11.6193C21.8195 11.6389 22.1626 11.6618 22.4429 11.6618V7.40825C22.3209 7.40825 22.1219 7.39596 21.7544 7.37149C21.4202 7.34925 20.9976 7.32115 20.5371 7.30672C19.6286 7.27824 18.4672 7.29779 17.3093 7.55377C16.1512 7.8098 14.8404 8.33724 13.8181 9.4452C12.7612 10.5907 12.2266 12.1236 12.2266 13.9619V15.0127H10.6836V19.2662H12.2266V26.6332H16.4801V19.2662H20.3394V15.0127H16.4801V13.9619Z" fill="currentColor"/></svg>save_model</span></h4><!-- HTML_TAG_END --> <a id="trl.DPOTrainer.save_model" class="header-link invisible with-hover:group-hover:visible pr-2" href="#trl.DPOTrainer.save_model"><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></a> <a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/trl/blob/vr_5607/transformers/trainer.py#L3746" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span class="hidden md:block mx-0.5 hover:!underline" data-svelte-h="svelte-122apf4">source</span> <span data-svelte-h="svelte-x0xyl0">></span></a></span> <p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span data-svelte-h="svelte-8mvn6a">(</span> <span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">output_dir<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">_internal_call<span class="opacity-60">: bool = False</span></span> </span> <span data-svelte-h="svelte-1jq0pl7">)</span> </p> <div class="!mb-10 relative docstring-details "> </div></div> <p data-svelte-h="svelte-r8h4ov">Will save the model, so you can reload it using <code>from_pretrained()</code>.</p> <p data-svelte-h="svelte-1e6bius">Will only save from the main process.</p></div> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"> <div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="trl.DPOTrainer.push_to_hub"><!-- HTML_TAG_START --><h4 class="!m-0"><span class="flex-1 rounded-xl py-0.5 break-all bg-gradient-to-r from-blue-50/60 to-white dark:from-gray-900 dark:to-gray-950 text-blue-700 dark:text-blue-300 font-medium px-2"><svg width="1em" height="1em" viewBox="0 0 32 33" class="mr-1 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg"><path d="M5.80566 18.3545C4.90766 17.4565 4.90766 16.0005 5.80566 15.1025L14.3768 6.53142C15.2748 5.63342 16.7307 5.63342 17.6287 6.53142L26.1999 15.1025C27.0979 16.0005 27.0979 17.4565 26.1999 18.3545L17.6287 26.9256C16.7307 27.8236 15.2748 27.8236 14.3768 26.9256L5.80566 18.3545Z" fill="currentColor" fill-opacity="0.25"/><path fill-rule="evenodd" clip-rule="evenodd" d="M16.4801 13.9619C16.4801 12.9761 16.7467 12.5436 16.9443 12.3296C17.1764 12.078 17.5731 11.8517 18.2275 11.707C18.8821 11.5623 19.638 11.5342 20.4038 11.5582C20.7804 11.57 21.1341 11.5932 21.4719 11.6156L21.5263 11.6193C21.8195 11.6389 22.1626 11.6618 22.4429 11.6618V7.40825C22.3209 7.40825 22.1219 7.39596 21.7544 7.37149C21.4202 7.34925 20.9976 7.32115 20.5371 7.30672C19.6286 7.27824 18.4672 7.29779 17.3093 7.55377C16.1512 7.8098 14.8404 8.33724 13.8181 9.4452C12.7612 10.5907 12.2266 12.1236 12.2266 13.9619V15.0127H10.6836V19.2662H12.2266V26.6332H16.4801V19.2662H20.3394V15.0127H16.4801V13.9619Z" fill="currentColor"/></svg>push_to_hub</span></h4><!-- HTML_TAG_END --> <a id="trl.DPOTrainer.push_to_hub" class="header-link invisible with-hover:group-hover:visible pr-2" href="#trl.DPOTrainer.push_to_hub"><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></a> <a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/trl/blob/vr_5607/transformers/trainer.py#L3993" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span class="hidden md:block mx-0.5 hover:!underline" data-svelte-h="svelte-122apf4">source</span> <span data-svelte-h="svelte-x0xyl0">></span></a></span> <p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span data-svelte-h="svelte-8mvn6a">(</span> <span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">commit_message<span class="opacity-60">: str | None = 'End of training'</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">blocking<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">token<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-pointer"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">revision<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">**kwargs<span class="opacity-60"></span></span> </span> <span data-svelte-h="svelte-1jq0pl7">)</span> </p> <div class="!mb-10 relative docstring-details "> <p class="flex items-center font-semibold !mt-2 !mb-2 text-gray-800" data-svelte-h="svelte-lt6pb6">Parameters <span class="flex-auto border-t-2 border-gray-100 dark:border-gray-700 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.push_to_hub.commit_message" class="header-link block pr-0.5 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="#trl.DPOTrainer.push_to_hub.commit_message"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>commit_message</strong> (<code>str</code>, <em>optional</em>, defaults to <code>"End of training"</code>) — | |
| Message to commit while pushing.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.push_to_hub.blocking" class="header-link block pr-0.5 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="#trl.DPOTrainer.push_to_hub.blocking"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>blocking</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) — | |
| Whether the function should return only when the <code>git push</code> has finished.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.push_to_hub.token" class="header-link block pr-0.5 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="#trl.DPOTrainer.push_to_hub.token"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>token</strong> (<code>str</code>, <em>optional</em>, defaults to <code>None</code>) — | |
| Token with write permission to overwrite Trainer’s original args.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.push_to_hub.revision" class="header-link block pr-0.5 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="#trl.DPOTrainer.push_to_hub.revision"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>revision</strong> (<code>str</code>, <em>optional</em>) — | |
| The git revision to commit from. Defaults to the head of the “main” branch.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOTrainer.push_to_hub.kwargs" class="header-link block pr-0.5 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="#trl.DPOTrainer.push_to_hub.kwargs"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>kwargs</strong> (<code>dict[str, Any]</code>, <em>optional</em>) — | |
| Additional keyword arguments passed along to <code>~Trainer.create_model_card</code>.<!-- HTML_TAG_END --> </span></span> </li></ul> </div></div> <p data-svelte-h="svelte-8tudwd">Upload <code>self.model</code> and <code>self.processing_class</code> to the 🤗 model hub on the repo <code>self.args.hub_model_id</code>.</p></div></div> <h2 class="relative group"><a id="trl.DPOConfig" 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="#trl.DPOConfig"><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>DPOConfig</span></h2> <div class="docstring border-l-2 border-t-2 pl-4 pt-3.5 border-gray-100 rounded-tl-xl mb-6 mt-8"> <div><span class="group flex space-x-1.5 items-center text-gray-800 bg-gradient-to-r rounded-tr-lg -mt-4 -ml-4 pt-3 px-2.5" id="trl.DPOConfig"><!-- HTML_TAG_START --><h3 class="!m-0"><span class="flex-1 break-all md:text-lg bg-gradient-to-r px-2.5 py-1.5 rounded-xl from-indigo-50/70 to-white dark:from-gray-900 dark:to-gray-950 dark:text-indigo-300 text-indigo-700"><svg class="mr-1.5 text-indigo-500 inline-block -mt-0.5" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" focusable="false" role="img" width=".8em" height=".8em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 24 24"><path class="uim-quaternary" d="M20.23 7.24L12 12L3.77 7.24a1.98 1.98 0 0 1 .7-.71L11 2.76c.62-.35 1.38-.35 2 0l6.53 3.77c.29.173.531.418.7.71z" opacity=".25" fill="currentColor"></path><path class="uim-tertiary" d="M12 12v9.5a2.09 2.09 0 0 1-.91-.21L4.5 17.48a2.003 2.003 0 0 1-1-1.73v-7.5a2.06 2.06 0 0 1 .27-1.01L12 12z" opacity=".5" fill="currentColor"></path><path class="uim-primary" d="M20.5 8.25v7.5a2.003 2.003 0 0 1-1 1.73l-6.62 3.82c-.275.13-.576.198-.88.2V12l8.23-4.76c.175.308.268.656.27 1.01z" fill="currentColor"></path></svg><span class="font-light">class</span> <span class="font-medium">trl.</span><span class="font-semibold">DPOConfig</span></span></h3><!-- HTML_TAG_END --> <a id="trl.DPOConfig" class="header-link invisible with-hover:group-hover:visible pr-2" href="#trl.DPOConfig"><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></a> <a class="!ml-auto !text-gray-400 !no-underline text-sm flex items-center" href="https://github.com/huggingface/trl/blob/vr_5607/trl/trainer/dpo_config.py#L23" target="_blank"><span data-svelte-h="svelte-1kd6by1"><</span> <span class="hidden md:block mx-0.5 hover:!underline" data-svelte-h="svelte-122apf4">source</span> <span data-svelte-h="svelte-x0xyl0">></span></a></span> <p class="font-mono text-xs md:text-sm !leading-relaxed !my-6"><span data-svelte-h="svelte-8mvn6a">(</span> <span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">output_dir<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">per_device_train_batch_size<span class="opacity-60">: int = 8</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">num_train_epochs<span class="opacity-60">: float = 3.0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">max_steps<span class="opacity-60">: int = -1</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">learning_rate<span class="opacity-60">: float = 1e-06</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">lr_scheduler_type<span class="opacity-60">: transformers.trainer_utils.SchedulerType | str = 'linear'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">lr_scheduler_kwargs<span class="opacity-60">: dict | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">warmup_steps<span class="opacity-60">: float = 0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">optim<span class="opacity-60">: transformers.training_args.OptimizerNames | str = 'adamw_torch_fused'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">optim_args<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">weight_decay<span class="opacity-60">: float = 0.0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">adam_beta1<span class="opacity-60">: float = 0.9</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">adam_beta2<span class="opacity-60">: float = 0.999</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">adam_epsilon<span class="opacity-60">: float = 1e-08</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">optim_target_modules<span class="opacity-60">: None | str | list[str] = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">gradient_accumulation_steps<span class="opacity-60">: int = 1</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">average_tokens_across_devices<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">max_grad_norm<span class="opacity-60">: float = 1.0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">label_smoothing_factor<span class="opacity-60">: float = 0.0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">bf16<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">fp16<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">bf16_full_eval<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">fp16_full_eval<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">tf32<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">gradient_checkpointing<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">gradient_checkpointing_kwargs<span class="opacity-60">: dict[str, typing.Any] | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">torch_compile<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">torch_compile_backend<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">torch_compile_mode<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">use_liger_kernel<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">liger_kernel_config<span class="opacity-60">: dict[str, bool] | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">use_cache<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">neftune_noise_alpha<span class="opacity-60">: float | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">torch_empty_cache_steps<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">auto_find_batch_size<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">logging_strategy<span class="opacity-60">: transformers.trainer_utils.IntervalStrategy | str = 'steps'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">logging_steps<span class="opacity-60">: float = 10</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">logging_first_step<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">log_on_each_node<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">logging_nan_inf_filter<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">include_num_input_tokens_seen<span class="opacity-60">: str | bool = 'no'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">log_level<span class="opacity-60">: str = 'passive'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">log_level_replica<span class="opacity-60">: str = 'warning'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">disable_tqdm<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">report_to<span class="opacity-60">: None | str | list[str] = 'none'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">run_name<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">project<span class="opacity-60">: str = 'huggingface'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">trackio_space_id<span class="opacity-60">: str | None = 'trackio'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_strategy<span class="opacity-60">: transformers.trainer_utils.IntervalStrategy | str = 'no'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_steps<span class="opacity-60">: float | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_delay<span class="opacity-60">: float = 0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">per_device_eval_batch_size<span class="opacity-60">: int = 8</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">prediction_loss_only<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_on_start<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_do_concat_batches<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_use_gather_object<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">eval_accumulation_steps<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">include_for_metrics<span class="opacity-60">: list = <factory></span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">batch_eval_metrics<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">save_only_model<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">save_strategy<span class="opacity-60">: transformers.trainer_utils.SaveStrategy | str = 'steps'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">save_steps<span class="opacity-60">: float = 500</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">save_on_each_node<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">save_total_limit<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">enable_jit_checkpoint<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">push_to_hub<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">hub_token<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">hub_private_repo<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">hub_model_id<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">hub_strategy<span class="opacity-60">: transformers.trainer_utils.HubStrategy | str = 'every_save'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">hub_always_push<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">hub_revision<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">load_best_model_at_end<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">metric_for_best_model<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">greater_is_better<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ignore_data_skip<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">restore_callback_states_from_checkpoint<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">full_determinism<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">seed<span class="opacity-60">: int = 42</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">data_seed<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">use_cpu<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">accelerator_config<span class="opacity-60">: dict | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">parallelism_config<span class="opacity-60">: accelerate.parallelism_config.ParallelismConfig | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">dataloader_drop_last<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">dataloader_num_workers<span class="opacity-60">: int = 0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">dataloader_pin_memory<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">dataloader_persistent_workers<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">dataloader_prefetch_factor<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">remove_unused_columns<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">label_names<span class="opacity-60">: list[str] | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">train_sampling_strategy<span class="opacity-60">: str = 'random'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">length_column_name<span class="opacity-60">: str = 'length'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ddp_find_unused_parameters<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ddp_bucket_cap_mb<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ddp_broadcast_buffers<span class="opacity-60">: bool | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ddp_backend<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ddp_timeout<span class="opacity-60">: int = 1800</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">fsdp<span class="opacity-60">: list[transformers.trainer_utils.FSDPOption] | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">fsdp_config<span class="opacity-60">: dict[str, typing.Any] | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">deepspeed<span class="opacity-60">: dict | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">debug<span class="opacity-60">: str | list[transformers.debug_utils.DebugOption] = ''</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">skip_memory_metrics<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">do_train<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">do_eval<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">do_predict<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">resume_from_checkpoint<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">warmup_ratio<span class="opacity-60">: float | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">logging_dir<span class="opacity-60">: str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">local_rank<span class="opacity-60">: int = -1</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">model_init_kwargs<span class="opacity-60">: dict[str, typing.Any] | str | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">disable_dropout<span class="opacity-60">: bool = True</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">dataset_num_proc<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">max_length<span class="opacity-60">: int | None = 1024</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">truncation_mode<span class="opacity-60">: str = 'keep_start'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">padding_free<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">pad_to_multiple_of<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">precompute_ref_log_probs<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">precompute_ref_batch_size<span class="opacity-60">: int | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">loss_type<span class="opacity-60">: list = <factory></span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">loss_weights<span class="opacity-60">: list[float] | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ld_alpha<span class="opacity-60">: float | None = None</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">f_divergence_type<span class="opacity-60">: str = 'reverse_kl'</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">f_alpha_divergence_coef<span class="opacity-60">: float = 0.5</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">label_smoothing<span class="opacity-60">: float = 0.0</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">beta<span class="opacity-60">: float = 0.1</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">use_weighting<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">discopop_tau<span class="opacity-60">: float = 0.05</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">activation_offloading<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">sync_ref_model<span class="opacity-60">: bool = False</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ref_model_mixup_alpha<span class="opacity-60">: float = 0.6</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">ref_model_sync_steps<span class="opacity-60">: int = 512</span></span> </span><span class="comma cursor-default"><span class="rounded hover:bg-black hover:text-white dark:hover:bg-white dark:hover:text-black">pad_token<span class="opacity-60">: str | None = None</span></span> </span> <span data-svelte-h="svelte-1jq0pl7">)</span> </p> <div class="!mb-10 relative docstring-details "> <p class="flex items-center font-semibold">Parameters that control the model <span class="flex-auto border-t-2 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.model_init_kwargs" class="header-link block pr-0.5 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="#trl.DPOConfig.model_init_kwargs"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>model_init_kwargs</strong> (<code>dict[str, Any]</code>, <em>optional</em>) — | |
| Keyword arguments for <a href="https://huggingface.co/docs/transformers/main/en/model_doc/auto#transformers.AutoModelForCausalLM.from_pretrained" rel="nofollow">from_pretrained</a>, used when the <code>model</code> | |
| argument of the <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a> is provided as a string.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.disable_dropout" class="header-link block pr-0.5 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="#trl.DPOConfig.disable_dropout"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>disable_dropout</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>True</code>) — | |
| Whether to disable dropout in the model and reference model.<!-- HTML_TAG_END --> </span></span> </li> </ul><p class="flex items-center font-semibold">Parameters that control the data preprocessing <span class="flex-auto border-t-2 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.dataset_num_proc" class="header-link block pr-0.5 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="#trl.DPOConfig.dataset_num_proc"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>dataset_num_proc</strong> (<code>int</code>, <em>optional</em>) — | |
| Number of processes to use for processing the dataset.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.max_length" class="header-link block pr-0.5 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="#trl.DPOConfig.max_length"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>max_length</strong> (<code>int</code> or <code>None</code>, <em>optional</em>, defaults to <code>1024</code>) — | |
| Maximum length of the tokenized sequence. Sequences longer than <code>max_length</code> are truncated from the left or | |
| right depending on the <code>truncation_mode</code>. If <code>None</code>, no truncation is applied.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.truncation_mode" class="header-link block pr-0.5 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="#trl.DPOConfig.truncation_mode"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>truncation_mode</strong> (<code>str</code>, <em>optional</em>, defaults to <code>"keep_start"</code>) — | |
| Truncation mode to use when the sequence exceeds <code>max_length</code>. The only supported value is | |
| <code>"keep_start"</code>. The <code>"keep_end"</code> value is deprecated and will be removed in v2.0.0.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.padding_free" class="header-link block pr-0.5 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="#trl.DPOConfig.padding_free"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>padding_free</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) — | |
| Whether to perform forward passes without padding by flattening all sequences in the batch into a single | |
| continuous sequence. This reduces memory usage by eliminating padding overhead. Currently, this is only | |
| supported with the FlashAttention 2 or 3, which can efficiently handle the flattened batch structure.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.pad_to_multiple_of" class="header-link block pr-0.5 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="#trl.DPOConfig.pad_to_multiple_of"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>pad_to_multiple_of</strong> (<code>int</code>, <em>optional</em>) — | |
| If set, the sequences will be padded to a multiple of this value.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.precompute_ref_log_probs" class="header-link block pr-0.5 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="#trl.DPOConfig.precompute_ref_log_probs"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>precompute_ref_log_probs</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) — | |
| Whether to precompute the reference model log probabilities for the entire training dataset before | |
| training. This allows to save memory during training, as the reference model does not need to be kept in | |
| memory.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.precompute_ref_batch_size" class="header-link block pr-0.5 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="#trl.DPOConfig.precompute_ref_batch_size"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>precompute_ref_batch_size</strong> (<code>int</code>, <em>optional</em>) — | |
| Batch size to use when precomputing reference model log probabilities. This can be set higher than the | |
| training batch size to speed up preprocessing. If <code>None</code>, defaults to <code>per_device_train_batch_size</code> for | |
| training and <code>per_device_eval_batch_size</code> for evaluation.<!-- HTML_TAG_END --> </span></span> </li> </ul><p class="flex items-center font-semibold">Parameters that control the training <span class="flex-auto border-t-2 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.loss_type" class="header-link block pr-0.5 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="#trl.DPOConfig.loss_type"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>loss_type</strong> (<code>list[str]</code>, <em>optional</em>, defaults to <code>["sigmoid"]</code>) — | |
| Type of loss to use. Possible values are: <code>'sigmoid'</code>, <code>'hinge'</code>, <code>'ipo'</code>, <code>'exo_pair'</code>, <code>'nca_pair'</code>, | |
| <code>'robust'</code>, <code>'bco_pair'</code>, <code>'sppo_hard'</code>, <code>'aot'</code>, <code>'aot_unpaired'</code>, <code>'apo_zero'</code>, <code>'apo_down'</code>, | |
| <code>'discopop'</code>, <code>'sft'</code>. If multiple loss types are provided, they will be combined using the weights | |
| specified in <code>loss_weights</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.loss_weights" class="header-link block pr-0.5 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="#trl.DPOConfig.loss_weights"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>loss_weights</strong> (<code>list[float]</code>, <em>optional</em>) — | |
| List of loss weights for multi-loss combinations. Used when combining multiple loss types. Example: <code>[0.8, 0.2, 1.0]</code> for MPO. If not provided, defaults to equal weights (<code>1.0</code>) for all loss types.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.ld_alpha" class="header-link block pr-0.5 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="#trl.DPOConfig.ld_alpha"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>ld_alpha</strong> (<code>float</code>, <em>optional</em>) — | |
| α parameter from the LD-DPO paper, which controls the weighting of the verbose token log-probabilities in | |
| responses. If <code>None</code>, no weighting is applied to the verbose part, and the loss is equivalent to the | |
| standard DPO loss. Must be in [0.0, 1.0]: <code>ld_alpha=1.0</code> applies no weighting, and <code>ld_alpha=0.0</code> masks | |
| tokens beyond shared lengths.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.f_divergence_type" class="header-link block pr-0.5 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="#trl.DPOConfig.f_divergence_type"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>f_divergence_type</strong> (<code>str</code>, <em>optional</em>, defaults to <code>"reverse_kl"</code>) — | |
| f-divergence regularizer between policy and reference (f-DPO paper). Possible values are: <code>reverse_kl</code> | |
| (default), <code>forward_kl</code>, <code>js_divergence</code>, <code>alpha_divergence</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.f_alpha_divergence_coef" class="header-link block pr-0.5 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="#trl.DPOConfig.f_alpha_divergence_coef"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>f_alpha_divergence_coef</strong> (<code>float</code>, <em>optional</em>, defaults to <code>0.5</code>) — | |
| α coefficient for the α-divergence u^-α regularizer, used only when <code>f_divergence_type='alpha_divergence'</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.label_smoothing" class="header-link block pr-0.5 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="#trl.DPOConfig.label_smoothing"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>label_smoothing</strong> (<code>float</code>, <em>optional</em>, defaults to <code>0.0</code>) — | |
| Label smoothing parameter used in Robust DPO and EXO. In Robust DPO, it is interpreted as the probability | |
| that a preference label is flipped and must lie in [0.0, 0.5); a typical value recommended by the Robust | |
| DPO paper is 0.1. In EXO, it corresponds to the ε label smoothing parameter, for which the paper recommends | |
| a typical value of 1e-3.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.beta" class="header-link block pr-0.5 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="#trl.DPOConfig.beta"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>beta</strong> (<code>float</code>, <em>optional</em>, defaults to <code>0.1</code>) — | |
| Parameter controlling the deviation from the reference model. Higher β means less deviation from the | |
| reference model. For the IPO loss (<code>loss_type='ipo'</code>), this value is the regularization parameter denoted | |
| by τ in the <a href="https://huggingface.co/papers/2310.12036" rel="nofollow">paper</a>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.use_weighting" class="header-link block pr-0.5 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="#trl.DPOConfig.use_weighting"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>use_weighting</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) — | |
| Whether to apply WPO-style weighting (<a href="https://huggingface.co/papers/2406.11827" rel="nofollow">https://huggingface.co/papers/2406.11827</a>) to preference pairs using | |
| the policy’s length-normalized sequence probabilities.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.discopop_tau" class="header-link block pr-0.5 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="#trl.DPOConfig.discopop_tau"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>discopop_tau</strong> (<code>float</code>, <em>optional</em>, defaults to <code>0.05</code>) — | |
| τ/temperature parameter from the DiscoPOP paper, which controls the shape of the log-ratio modulated loss | |
| when using <code>loss_type='discopop'</code>. The paper recommends the default value <code>discopop_tau=0.05</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.activation_offloading" class="header-link block pr-0.5 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="#trl.DPOConfig.activation_offloading"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>activation_offloading</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) — | |
| Whether to offload the activations to the CPU.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.sync_ref_model" class="header-link block pr-0.5 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="#trl.DPOConfig.sync_ref_model"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>sync_ref_model</strong> (<code>bool</code>, <em>optional</em>, defaults to <code>False</code>) — | |
| Whether to synchronize the reference model with the active model every <code>ref_model_sync_steps</code> steps, using | |
| the <code>ref_model_mixup_alpha</code> parameter. This synchronization originates from the | |
| <a href="https://huggingface.co/papers/2404.09656" rel="nofollow">TR-DPO</a> paper. <code>sync_ref_model=True</code> is not yet compatible with | |
| PEFT or <code>precompute_ref_log_probs=True</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.ref_model_mixup_alpha" class="header-link block pr-0.5 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="#trl.DPOConfig.ref_model_mixup_alpha"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>ref_model_mixup_alpha</strong> (<code>float</code>, <em>optional</em>, defaults to <code>0.6</code>) — | |
| α parameter from the TR-DPO paper, which controls the mix between the current policy and the previous | |
| reference policy during updates. The reference policy is updated according to the equation: <code>π_ref = α * π_θ + (1 - α) * π_ref_prev</code>. To use this parameter, you must set <code>sync_ref_model=True</code>.<!-- HTML_TAG_END --> </span></span> </li><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.ref_model_sync_steps" class="header-link block pr-0.5 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="#trl.DPOConfig.ref_model_sync_steps"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>ref_model_sync_steps</strong> (<code>int</code>, <em>optional</em>, defaults to <code>512</code>) — | |
| τ parameter from the TR-DPO paper, which determines how frequently the current policy is synchronized with | |
| the reference policy. To use this parameter, you must set <code>sync_ref_model=True</code>.<!-- HTML_TAG_END --> </span></span> </li> </ul><p class="flex items-center font-semibold">Deprecated parameters <span class="flex-auto border-t-2 ml-3"></span></p> <ul class="px-2"><li class="text-base !pl-4 my-3 rounded "><span class="group flex space-x-1.5 items-start"><a id="trl.DPOConfig.pad_token" class="header-link block pr-0.5 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="#trl.DPOConfig.pad_token"><span><svg class="text-smd" 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><!-- HTML_TAG_START --><strong>pad_token</strong> —</p> | |
| <deprecated version="1.1.0"> | |
| <p>Parameter <code>pad_token</code> is deprecated and will be removed in version v2.0.0. Set <code>tokenizer.pad_token</code> | |
| directly and pass it as <code>processing_class</code> to the trainer instead.</p> | |
| </deprecated><!-- HTML_TAG_END --> </span></span> </li> </ul> </div></div> <p data-svelte-h="svelte-1yv7gy7">Configuration class for the <a href="/docs/trl/pr_5607/en/bema_for_reference_model#trl.DPOTrainer">DPOTrainer</a>.</p> <p data-svelte-h="svelte-1p9pbif">This class includes only the parameters that are specific to DPO training. For a full list of training arguments, | |
| please refer to the <a href="https://huggingface.co/docs/transformers/main/en/main_classes/trainer#transformers.TrainingArguments" rel="nofollow">TrainingArguments</a> documentation. Note that default values in this class may | |
| differ from those in <a href="https://huggingface.co/docs/transformers/main/en/main_classes/trainer#transformers.TrainingArguments" rel="nofollow">TrainingArguments</a>.</p> <p data-svelte-h="svelte-ekuf1t">Using <a href="https://huggingface.co/docs/transformers/main/en/internal/trainer_utils#transformers.HfArgumentParser" rel="nofollow">HfArgumentParser</a> we can turn this class into | |
| <a href="https://docs.python.org/3/library/argparse#module-argparse" rel="nofollow">argparse</a> arguments that can be specified on the | |
| command line.</p> <blockquote class="note" data-svelte-h="svelte-17fyuhe"><p>These parameters have default values different from <a href="https://huggingface.co/docs/transformers/main/en/main_classes/trainer#transformers.TrainingArguments" rel="nofollow">TrainingArguments</a>:</p> <ul><li><code>logging_steps</code>: Defaults to <code>10</code> instead of <code>500</code>.</li> <li><code>gradient_checkpointing</code>: Defaults to <code>True</code> instead of <code>False</code>.</li> <li><code>bf16</code>: Defaults to <code>True</code> if <code>fp16</code> is not set, instead of <code>False</code>.</li> <li><code>learning_rate</code>: Defaults to <code>1e-6</code> instead of <code>5e-5</code>.</li></ul></blockquote></div> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/trl/blob/main/docs/source/dpo_trainer.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 data-svelte-h="svelte-zjs2n5"><span class="underline">Update</span> on GitHub</span></a> <p></p> | |
| <script> | |
| { | |
| __sveltekit_1hqaf25 = { | |
| assets: "/docs/trl/pr_5607/en", | |
| base: "/docs/trl/pr_5607/en", | |
| env: {} | |
| }; | |
| const element = document.currentScript.parentElement; | |
| const data = [null,null]; | |
| Promise.all([ | |
| import("/docs/trl/pr_5607/en/_app/immutable/entry/start.151d81bd.js"), | |
| import("/docs/trl/pr_5607/en/_app/immutable/entry/app.3d9a91c0.js") | |
| ]).then(([kit, app]) => { | |
| kit.start(app, element, { | |
| node_ids: [0, 16], | |
| data, | |
| form: null, | |
| error: null | |
| }); | |
| }); | |
| } | |
| </script> | |
Xet Storage Details
- Size:
- 232 kB
- Xet hash:
- 5bc4b45857f17e711f72b8294b9b5921b5d48cfb6696eca41d976b2d8993fb19
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.