qwen2.5-0.5b-anghabench-32kcw / train_results.json
ahmedheakl's picture
End of training
f072299 verified
{
"epoch": 1.0,
"total_flos": 389042235549696.0,
"train_loss": 0.008355458706034012,
"train_runtime": 62709.2593,
"train_samples_per_second": 7.797,
"train_steps_per_second": 0.975
}