Posts
From the last year
trl version >0.23.0 breaks with minimal dependencies due to wandb Weave unconditional import
python trl versioning breaking-change wandb 209 tokens
When training Gemma 4 (4B or 31B variants) using HuggingFace's
DPOTrainer with text-only prompt/chosen/rejected triples, training fails immediately with:python gemma huggingface trl dpo-trainer 114 tokens