qwen2-math-7b-step-dpo / all_results.json
rasdani's picture
Model save
134434e verified
raw
history blame contribute delete
No virus
232 Bytes
{
"epoch": 7.964444444444444,
"total_flos": 0.0,
"train_loss": 0.11349717956385402,
"train_runtime": 13006.5105,
"train_samples": 10795,
"train_samples_per_second": 6.64,
"train_steps_per_second": 0.103
}