Files
Ministral-8B-Instruct-2410_…/train_results.json

8 lines
220 B
JSON
Raw Normal View History

{
"total_flos": 5.7509457374759354e+19,
"train_loss": 0.3138526783479623,
"train_runtime": 47000.1043,
"train_samples": 125770,
"train_samples_per_second": 8.028,
"train_steps_per_second": 0.502
}