Qwen2.5-1.5B-hl-false-v4 / train_results.json
Muennighoff's picture
Model save
069ef98 verified
{
"total_flos": 0.0,
"train_loss": -0.03081917177627043,
"train_runtime": 101643.0755,
"train_samples": 12000,
"train_samples_per_second": 14.104,
"train_steps_per_second": 0.016
}