math_llama3_reset_dpo_100_0_1.0 / train_results.json
lzc0525's picture
Upload folder using huggingface_hub
9344507 verified
{
"epoch": 0.9221902017291066,
"total_flos": 0.0,
"train_loss": 0.45325408279895785,
"train_runtime": 617.7917,
"train_samples": 1387,
"train_samples_per_second": 2.245,
"train_steps_per_second": 0.016
}