Qwen2.5-1.5B-Open-R1-Distill / train_results.json
YWZBrandon's picture
Training in progress, step 100
013e29b verified
raw
history blame
214 Bytes
{
"total_flos": 489077052801024.0,
"train_loss": 0.5497076121425362,
"train_runtime": 14506.5548,
"train_samples": 93733,
"train_samples_per_second": 2.365,
"train_steps_per_second": 0.074
}