qwen2.5-3b-ArabicSFT_1epoch / all_results.json
RedaAlami's picture
Initial model upload
c310bd2 verified
{
"total_flos": 69610707615744.0,
"train_loss": 0.6555303625158362,
"train_runtime": 3309.6115,
"train_samples": 3533,
"train_samples_per_second": 1.067,
"train_steps_per_second": 0.034
}