e1_math_all_r1 / train_results.json
ryanmarten's picture
End of training
5a8e2a5 verified
raw
history blame contribute delete
221 Bytes
{
"epoch": 4.982278481012658,
"total_flos": 5.951770096669229e+18,
"train_loss": 0.3465574411841912,
"train_runtime": 60944.3549,
"train_samples_per_second": 2.593,
"train_steps_per_second": 0.02
}