zilo-instruct-v2-sft-qlora / all_results.json
Mark-Arcee's picture
End of training
b4f744e verified
raw
history blame
No virus
414 Bytes
{
"epoch": 2.994174757281553,
"eval_loss": 0.38681623339653015,
"eval_runtime": 22.0502,
"eval_samples": 1530,
"eval_samples_per_second": 4.989,
"eval_steps_per_second": 0.635,
"total_flos": 5.421128640186286e+17,
"train_loss": 0.2601610358741651,
"train_runtime": 3322.581,
"train_samples": 29084,
"train_samples_per_second": 1.86,
"train_steps_per_second": 0.232
}