llama3-1_8b_math_100000_samples / train_results.json
gsmyrnis's picture
End of training
2dc6861 verified
raw
history blame contribute delete
204 Bytes
{
"epoch": 3.0,
"total_flos": 422109385850880.0,
"train_loss": 0.5631871157222323,
"train_runtime": 1874.1971,
"train_samples_per_second": 68.655,
"train_steps_per_second": 0.134
}