Llama-3.1-8B-coding / train_results.json
pmahdavi's picture
Upload coding model (without optimizer states)
06a3c1b verified
raw
history blame contribute delete
218 Bytes
{
"epoch": 0.9995220557226798,
"total_flos": 145966987345920.0,
"train_loss": 0.845390060482317,
"train_runtime": 63839.4451,
"train_samples_per_second": 2.229,
"train_steps_per_second": 0.017
}