llama3.2-1b-deita-dpo-student_sft_init/train_results.json

{
    "epoch": 3.0,
    "total_flos": 90953314467840.0,
    "train_loss": 0.9790556504583052,
    "train_runtime": 12705.6497,
    "train_samples": 9500,
    "train_samples_per_second": 8.618,
    "train_steps_per_second": 0.135
}