mix_base_code_math_t_1 / train_results.json
sedrickkeh's picture
End of training
beba1cb verified
{
"epoch": 2.9952,
"total_flos": 312484367761408.0,
"train_loss": 0.6120152068443787,
"train_runtime": 29530.8856,
"train_samples_per_second": 1.016,
"train_steps_per_second": 0.011
}