tinyllama_mole_sft_router05_ep3 / train_results.json
hushell's picture
Model save
729e41c verified
{
"epoch": 3.0,
"train_loss": 2.0904923673613602,
"train_runtime": 47734.51,
"train_samples": 207865,
"train_samples_per_second": 9.179,
"train_steps_per_second": 0.072
}