mamba_0_5_sft / train_results.json
Junxiong Wang
add models
74d28bb
{
"epoch": 1.0,
"total_flos": 7.680275121596006e+16,
"train_loss": 0.6780993922750357,
"train_runtime": 284116.9477,
"train_samples": 15487525,
"train_samples_per_second": 12.403,
"train_steps_per_second": 0.194
}