smolm-autoreg-bpe-counterfactual_babylm_aann_high_variability_numeral-seed_211-1e-3
/
train_results.json
{ | |
"epoch": 20.0, | |
"train_loss": 3.0296430246833577, | |
"train_runtime": 81436.2617, | |
"train_samples": 595065, | |
"train_samples_per_second": 146.143, | |
"train_steps_per_second": 4.567 | |
} |