Qwen2.5-0.5B-Instruct_Short_CoT / train_results.json
TaiGary's picture
Add files using upload-large-folder tool
cf7bcf4 verified
raw
history blame contribute delete
202 Bytes
{
"epoch": 2.0,
"total_flos": 3605863612416.0,
"train_loss": 0.12988467803392226,
"train_runtime": 403.4892,
"train_samples_per_second": 26.415,
"train_steps_per_second": 1.656
}