math_llama3_reset_dpo_100_0_0.0 / train_results.json
lzc0525's picture
Upload folder using huggingface_hub
da085a6 verified
raw
history blame contribute delete
230 Bytes
{
"epoch": 0.9182209469153515,
"total_flos": 0.0,
"train_loss": 0.6934699952602387,
"train_runtime": 619.8127,
"train_samples": 1393,
"train_samples_per_second": 2.247,
"train_steps_per_second": 0.016
}