Files
qwen3-4b-refiner-gpt54-ep2/train_results.json

8 lines
202 B
JSON
Raw Permalink Normal View History

{
"epoch": 2.0,
"total_flos": 50054593118208.0,
"train_loss": 0.4503334106400956,
"train_runtime": 8072.0948,
"train_samples_per_second": 3.406,
"train_steps_per_second": 0.107
}