smolm-autoreg-bpe-counterfactual_babylm_aann_low_variability_numeral-seed_1024-1e-3
/
train_results.json
{ | |
"epoch": 20.0, | |
"train_loss": 3.027371880730273, | |
"train_runtime": 81125.0441, | |
"train_samples": 594966, | |
"train_samples_per_second": 146.679, | |
"train_steps_per_second": 4.584 | |
} |