b1_math_top_8_1k / train_results.json
ryanmarten's picture
Upload model
ef58e85 verified
{
"epoch": 6.72,
"total_flos": 1.0705425667824026e+17,
"train_loss": 0.4356019737465041,
"train_runtime": 4769.8365,
"train_samples_per_second": 1.468,
"train_steps_per_second": 0.015
}