openthoughts3_100k_qwen25_1b_bsz256_lr4e5_epochs7/train_results.json

{
    "epoch": 7.0,
    "total_flos": 9969287656374272.0,
    "train_loss": 1.0345919296651762,
    "train_runtime": 97654.46,
    "train_samples_per_second": 7.168,
    "train_steps_per_second": 0.028
}