Files
qwen2.5-3b-sft/all_results.json

8 lines
209 B
JSON
Raw Permalink Normal View History

{
"epoch": 3.0,
"total_flos": 6.08604739672526e+18,
"train_loss": 0.010436751544759131,
"train_runtime": 12813.4993,
"train_samples_per_second": 3.885,
"train_steps_per_second": 0.061
}