llama_2_llama_2_code_math_3_full / train_results.json
CharlesLi's picture
Model save
dfa2819 verified
{
"epoch": 1.0,
"total_flos": 2355521126400.0,
"train_loss": 0.8085391936094865,
"train_runtime": 126.1132,
"train_samples": 3980,
"train_samples_per_second": 5.804,
"train_steps_per_second": 0.182
}