ex19_qwen2.5-1.5b-1M-stack-16kcw / train_results.json
ahmedheakl's picture
End of training
70049bf verified
{
"epoch": 2.0,
"total_flos": 4010407409123328.0,
"train_loss": 0.0024581665951860818,
"train_runtime": 623670.1327,
"train_samples_per_second": 3.9,
"train_steps_per_second": 0.487
}