TRL documentation
快速入门
You are viewing main version, which requires installation from source. If you'd like
regular pip install, checkout the latest stable version (v1.17.1).
快速入门
TRL 是一个全面的库，用于使用监督微调（SFT）、组相对策略优化（GRPO）、直接偏好优化（DPO）等技术对基础模型进行后训练。
快速示例
立即使用 TRL 中最受欢迎的训练器开始上手。每个示例都使用紧凑的模型，以便快速进行实验。
监督微调
from trl import SFTTrainer
from datasets import load_dataset
trainer = SFTTrainer(
model="Qwen/Qwen2.5-0.5B",
train_dataset=load_dataset("trl-lib/Capybara", split="train"),
)
trainer.train()Group Relative Policy Optimization
accuracy_reward 需要 math-verify，您可以通过 pip install trl[math_verify] 进行安装。
from trl import GRPOTrainer
from datasets import load_dataset
from trl.rewards import accuracy_reward
trainer = GRPOTrainer(
model="Qwen/Qwen2.5-0.5B-Instruct",
train_dataset=load_dataset("trl-lib/DeepMath-103K", split="train"),
reward_funcs=accuracy_reward,
)
trainer.train()直接偏好优化
from trl import DPOTrainer
from datasets import load_dataset
trainer = DPOTrainer(
model="Qwen/Qwen2.5-0.5B-Instruct",
train_dataset=load_dataset("trl-lib/ultrafeedback_binarized", split="train"),
)
trainer.train()奖励建模
from trl import RewardTrainer
from datasets import load_dataset
dataset = load_dataset("trl-lib/ultrafeedback_binarized", split="train")
trainer = RewardTrainer(
model="Qwen/Qwen2.5-0.5B-Instruct",
train_dataset=dataset,
)
trainer.train()命令行界面
完全跳过代码 - 直接从终端开始训练：
# SFT: Fine-tune on instructions
trl sft --model_name_or_path Qwen/Qwen2.5-0.5B \
--dataset_name trl-lib/Capybara
# DPO: Align with preferences
trl dpo --model_name_or_path Qwen/Qwen2.5-0.5B-Instruct \
--dataset_name trl-lib/ultrafeedback_binarized
# Reward: Train a reward model
trl reward --model_name_or_path Qwen/Qwen2.5-0.5B-Instruct \
--dataset_name trl-lib/ultrafeedback_binarized接下来是什么？
📚 了解更多
- SFT Trainer - 完整的 SFT 指南
- DPO Trainer - 偏好对齐
- GRPO Trainer - 组相对策略优化
🚀 扩展
💡 示例
故障排除
内存不足？
减小批量大小并启用优化：
SFT
DPO
GRPO
training_args = SFTConfig(
per_device_train_batch_size=1, # Start small
gradient_accumulation_steps=8, # Maintain effective batch size
)training_args = DPOConfig(
per_device_train_batch_size=1, # Start small
gradient_accumulation_steps=8, # Maintain effective batch size
)training_args = GRPOConfig(
per_device_train_batch_size=1, # Start small
gradient_accumulation_steps=8, # Maintain effective batch size
num_generations=4, # Reduce from default 8 (GRPO generates num_generations completions per prompt)
max_completion_length=512, # Tune based on task; longer sequences cost more memory
)损失不下降？
尝试调整学习率：
training_args = SFTConfig(learning_rate=2e-5) # Good starting point如需更多帮助，请在 GitHub 上提交 issue。
Update on GitHub