Files
Qwen3-8B-SFT-envbench_gpt5-…/all_results.json

14 lines
426 B
JSON
Raw Permalink Normal View History

{
"epoch": 5.0,
"eval_accuracy": 0.8568963869568719,
"eval_loss": 0.4850449860095978,
"eval_runtime": 4.2558,
"eval_samples_per_second": 11.749,
"eval_steps_per_second": 3.055,
"num_input_tokens_seen": 31362728,
"total_flos": 1.4241950524474655e+18,
"train_loss": 0.4095978515633082,
"train_runtime": 3265.849,
"train_samples_per_second": 1.442,
"train_steps_per_second": 0.09
}