Files
Qwen3-1.7B-SFT-science-2e-5/all_results.json

12 lines
351 B
JSON
Raw Normal View History

{
"epoch": 3.0,
"eval_loss": 0.7464115023612976,
"eval_runtime": 140.9584,
"eval_samples_per_second": 36.834,
"eval_steps_per_second": 9.208,
"total_flos": 4.1917370482093916e+18,
"train_loss": 0.06833103949818527,
"train_runtime": 3477.4713,
"train_samples_per_second": 85.09,
"train_steps_per_second": 0.665
}