nemo_nano_math_300k / train_results.json
sedrickkeh's picture
End of training
16fe58b verified
raw
history blame contribute delete
225 Bytes
{
"epoch": 4.998573466476462,
"total_flos": 2.8177610658514207e+19,
"train_loss": 0.41344334662777105,
"train_runtime": 239120.7765,
"train_samples_per_second": 2.345,
"train_steps_per_second": 0.005
}