d1_code_all_large_10k / train_results.json
ryanmarten's picture
End of training
95661dd verified
{
"epoch": 4.992,
"total_flos": 2.1207819173211668e+18,
"train_loss": 0.5314857248312388,
"train_runtime": 63770.7246,
"train_samples_per_second": 0.784,
"train_steps_per_second": 0.006
}