Files
qwen3-8b-base-sft-hh-harmle…/all_results.json

14 lines
401 B
JSON
Raw Permalink Normal View History

{
"epoch": 1.0,
"eval_loss": 1.6253376007080078,
"eval_runtime": 2.3591,
"eval_samples": 2303,
"eval_samples_per_second": 330.634,
"eval_steps_per_second": 2.967,
"total_flos": 4.017624413372416e+16,
"train_loss": 1.907559284457454,
"train_runtime": 456.0106,
"train_samples": 42336,
"train_samples_per_second": 30.304,
"train_steps_per_second": 0.237
}