openthoughts3_100k_qwen25_1b_bsz256_lr2e5_epochs7/train_results.json

{
    "epoch": 7.0,
    "total_flos": 9969287656374272.0,
    "train_loss": 1.063590354692078,
    "train_runtime": 97730.0822,
    "train_samples_per_second": 7.163,
    "train_steps_per_second": 0.028
}