Qwen3-1.7B-debug / train_results.json
Muennighoff's picture
Model save
b5ea1b8 verified
{
"total_flos": 0.0,
"train_loss": 0.00037381448991557895,
"train_runtime": 2088.7955,
"train_samples": 40315,
"train_samples_per_second": 0.287,
"train_steps_per_second": 0.144
}