train / configs /Qwen3-4B-Base_a10g-small.yaml
qgallouedec's picture
qgallouedec HF Staff
demo config and run trl sft
c5d1a86
raw
history blame
68 Bytes
model_name_or_path: Qwen/Qwen2.5-0.5B
dataset_name: stanfordnlp/imdb