Files
Qwen2.5-7B-ODA-Mixture-500k/train_results.json

8 lines
210 B
JSON
Raw Normal View History

{
"epoch": 3.0,
"total_flos": 1.440651216420864e+16,
"train_loss": 0.30719185047274594,
"train_runtime": 200812.4774,
"train_samples_per_second": 1.094,
"train_steps_per_second": 0.034
}