recreate_llama_68M_vanilla / train_results.json
DorinSht's picture
End of training
5d9ef80 verified
raw
history blame contribute delete
238 Bytes
{
"epoch": 3.0,
"total_flos": 1.4536404559724544e+17,
"train_loss": 2.5941595100713495,
"train_runtime": 20556.3593,
"train_samples": 90745,
"train_samples_per_second": 13.243,
"train_steps_per_second": 0.552
}