Files
qwen3-4b-base-prompt/all_results.json

12 lines
353 B
JSON
Raw Normal View History

{
"epoch": 3.0,
"eval_loss": 0.008671932853758335,
"eval_runtime": 90.8179,
"eval_samples_per_second": 10.053,
"eval_steps_per_second": 5.032,
"total_flos": 5.550091697158881e+18,
"train_loss": 0.008026252662842327,
"train_runtime": 22347.8789,
"train_samples_per_second": 2.327,
"train_steps_per_second": 0.073
}