astrollama-3-8b-chat_summary / train_results.json
TirthankarSlg's picture
sft done
540a156 verified
raw
history blame
235 Bytes
{
"epoch": 3.0,
"total_flos": 6.976297471026659e+18,
"train_loss": 0.7863800849406247,
"train_runtime": 8818.5969,
"train_samples": 25188,
"train_samples_per_second": 8.569,
"train_steps_per_second": 0.134
}