b2_math_length_gpt41nano / train_results.json
neginr's picture
End of training
613f4e2 verified
raw
history blame contribute delete
208 Bytes
{
"epoch": 5.0,
"total_flos": 4.764433372703359e+18,
"train_loss": 0.3434084688603637,
"train_runtime": 29328.3301,
"train_samples_per_second": 5.386,
"train_steps_per_second": 0.042
}