smolm-autoreg-bpe-counterfactual_babylm_aann_high_variability_noun-seed_1024-1e-3
/
train_results.json
{ | |
"epoch": 20.0, | |
"total_flos": 1.56681778719744e+18, | |
"train_loss": 3.0269814196563325, | |
"train_runtime": 81729.3096, | |
"train_samples": 594994, | |
"train_samples_per_second": 145.601, | |
"train_steps_per_second": 4.55 | |
} |