Llama-3.1-8B-Instruct-SAA-100 / train_results.json
chchen's picture
End of training
08e2b67 verified
raw
history blame contribute delete
215 Bytes
{
"epoch": 8.88888888888889,
"total_flos": 9214586883145728.0,
"train_loss": 1.5806806755065919,
"train_runtime": 281.377,
"train_samples_per_second": 3.199,
"train_steps_per_second": 0.178
}