🏆Reinforcement Learning - DPO, ORPO & KTO

To use the reward modelling functions for DPO, ORPO or KTO with Unsloth, follow the steps below:

DPO (Direct Preference Optimization), ORPO (Odds Ratio Preference Optimization), PPO, KTO Reward Modelling all work with Unsloth.

We have Google Colab notebooks for reproducing ORPO, DPO Zephyr, KTO and SimPO:

• ORPO notebook

• DPO Zephyr notebook

• KTO notebook

• SimPO notebook

We're also in 🤗Hugging Face's official docs! We're on the SFT docs and the DPO docs.

DPO Code

python
import os
os.environ["CUDA_VISIBLE_DEVICES"] = "0" # Optional set GPU device ID

from unsloth import FastLanguageModel, PatchDPOTrainer
from unsloth import is_bfloat16_supported
PatchDPOTrainer()
import torch
from transformers import TrainingArguments
from trl import DPOTrainer

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

# Do model patching and add fast LoRA weights
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, # Supports any, but = 0 is optimized
    bias = "none",    # Supports any, but = "none" is optimized
    # [NEW] "unsloth" uses 30% less VRAM, fits 2x larger batch sizes!
    use_gradient_checkpointing = "unsloth", # True or "unsloth" for very long context
    random_state = 3407,
    max_seq_length = max_seq_length,
)

dpo_trainer = DPOTrainer(
    model = model,
    ref_model = None,
    args = TrainingArguments(
        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()