Llama8B_mathinstruct_SFT / train_results.json
Changahou's picture
Upload model
05dbd61 verified
Raw
History Blame Contribute Delete
248 Bytes
{
"epoch": 3.0,
"num_input_tokens_seen": 70167528,
"total_flos": 3.161046812140241e+18,
"train_loss": 0.5038315440246255,
"train_runtime": 10949.1572,
"train_samples_per_second": 27.399,
"train_steps_per_second": 0.571
}