Llama-3-Instruct-8B-CPO / train_results.json
haoranxu's picture
Upload folder using huggingface_hub
fea62aa verified
raw
history blame
217 Bytes
{
"epoch": 1.0,
"total_flos": 0.0,
"train_loss": 1.4382328207676227,
"train_runtime": 9608.0854,
"train_samples": 59876,
"train_samples_per_second": 6.232,
"train_steps_per_second": 0.049
}