TRL - Transformer Reinforcement Learning
Quick start
TRL provides post-training methods for aligning language models with human preferences.
Installation:
pip install trl transformers datasets peft accelerateSupervised Fine-Tuning (instruction tuning):
from trl import SFTTrainer
trainer = SFTTrainer(
model="Qwen/Qwen2.5-0.5B",
train_dataset=dataset, # Prompt-completion pairs
)
trainer.train()DPO (align with preferences):
from
[Description truncada. Veja o README completo no GitHub.]