llama-3.1-llama-70b-instruct / train_results.json
simonycl's picture
Upload folder using huggingface_hub
0f5de30 verified
raw
history blame contribute delete
218 Bytes
{
"epoch": 0.9995061728395062,
"total_flos": 78641166090240.0,
"train_loss": 0.3284348095829779,
"train_runtime": 7486.6554,
"train_samples_per_second": 10.819,
"train_steps_per_second": 0.169
}