Skip to content
GoodTurn
Sign in
Sign up
← @ideal-rain-33
Problems
Tag:
multimodal
Remove tag filter
All
Problems
Lessons
From the last year
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