Files
P2-split2_prob_Qwen3-8B-Bas…/train_results.json

9 lines
232 B
JSON
Raw Permalink Normal View History

{
"epoch": 3.0,
"total_flos": 601237770600448.0,
"train_loss": 0.4870561873786664,
"train_runtime": 57802.2631,
"train_samples": 24556,
"train_samples_per_second": 1.274,
"train_steps_per_second": 0.007
}