> For the complete documentation index, see [llms.txt](https://unsloth.ai/docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://unsloth.ai/docs/fr/commencer/reinforcement-learning-rl-guide/preference-dpo-orpo-and-kto.md).

# Entraînement de l'optimisation des préférences - DPO, ORPO et KTO

Découvrez le fine-tuning d'alignement des préférences avec DPO, GRPO, ORPO ou KTO via Unsloth, suivez les étapes ci-dessous :

DPO (Direct Preference Optimization), ORPO (Odds Ratio Preference Optimization), PPO et la modélisation de récompense KTO fonctionnent tous avec Unsloth.

Nous avons des notebooks Google Colab pour reproduire GRPO, ORPO, DPO Zephyr, KTO et SimPO :

* [Notebooks GRPO](/docs/fr/commencer/unsloth-notebooks.md#grpo-reasoning-rl-notebooks)
* [Notebook ORPO](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Llama3_\(8B\)-ORPO.ipynb)
* [Notebook DPO Zephyr](https://colab.research.google.com/github/unslothai/notebooks/blob/main/nb/Zephyr_\(7B\)-DPO.ipynb)
* [Notebook KTO](https://colab.research.google.com/drive/1MRgGtLWuZX4ypSfGguFgC-IblTvO2ivM?usp=sharing)
* [Notebook SimPO](https://colab.research.google.com/drive/1Hs5oQDovOay4mFA6Y9lQhVJ8TnbFLFh2?usp=sharing)

Nous sommes aussi dans les docs officielles de 🤗Hugging Face ! Nous sommes dans les [docs SFT](https://huggingface.co/docs/trl/main/en/sft_trainer#accelerate-fine-tuning-2x-using-unsloth) et les [docs DPO](https://huggingface.co/docs/trl/main/en/dpo_trainer#accelerate-dpo-fine-tuning-using-unsloth).

## Code DPO

```python
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "0" # Définition facultative de l'ID du périphérique GPU

from unsloth import FastLanguageModel, PatchDPOTrainer
from unsloth import is_bfloat16_supported
PatchDPOTrainer()
import torch
from trl import DPOTrainer, DPOConfig  # Modifié à partir de TrainingArguments

model, tokenizer = FastLanguageModel.from_pretrained(
    model_name = "unsloth/zephyr-sft-bnb-4bit",
    max_seq_length = max_seq_length,
    dtype = None,
    load_in_4bit = True,
)

# Effectuer le patch du modèle et ajouter des poids LoRA rapides
model = FastLanguageModel.get_peft_model(
    model,
    r = 64,
    target_modules = ["q_proj", "k_proj", "v_proj", "o_proj",
                      "gate_proj", "up_proj", "down_proj",],
    lora_alpha = 64,
    lora_dropout = 0, # Prend en charge n’importe quelle valeur, mais = 0 est optimisé
    bias = "none",    # Prend en charge n’importe quelle valeur, mais = "none" est optimisé
    # [NOUVEAU] "unsloth" utilise 30 % de VRAM en moins, permet des tailles de lot 2x plus grandes !
    use_gradient_checkpointing = "unsloth", # True ou "unsloth" pour un contexte très long
    random_state = 3407,
    max_seq_length = max_seq_length,
)

dpo_trainer = DPOTrainer(
    model = model,
    ref_model = None,
    args = DPOConfig( # Utiliser DPOConfig
        per_device_train_batch_size = 4,
        gradient_accumulation_steps = 8,
        warmup_ratio = 0.1,
        num_train_epochs = 3,
        fp16 = not is_bfloat16_supported(),
        bf16 = is_bfloat16_supported(),
        logging_steps = 1,
        optim = "adamw_8bit",
        seed = 42,
        output_dir = "outputs",
    ),
    beta = 0.1,
    train_dataset = YOUR_DATASET_HERE,
    # eval_dataset = YOUR_DATASET_HERE,
    tokenizer = tokenizer,
    max_length = 1024,
    max_prompt_length = 512,
)

dpo_trainer.train()
```


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://unsloth.ai/docs/fr/commencer/reinforcement-learning-rl-guide/preference-dpo-orpo-and-kto.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.