d1_code_all_1k / train_results.json
ryanmarten's picture
End of training
761ddbf verified
{
"epoch": 6.72,
"total_flos": 2.8410057313170227e+17,
"train_loss": 0.638406787599836,
"train_runtime": 8875.7256,
"train_samples_per_second": 0.789,
"train_steps_per_second": 0.008
}