Qwen-2.5-7B-Simple-RL / all_results.json
andrewsiah's picture
End of training
03635e6 verified
raw
history blame contribute delete
372 Bytes
{
"eval_loss": 0.00016938515182118863,
"eval_runtime": 4375.5256,
"eval_samples": 5000,
"eval_samples_per_second": 1.143,
"eval_steps_per_second": 0.01,
"total_flos": 0.0,
"train_loss": 0.0002975721653738533,
"train_runtime": 28855.0053,
"train_samples": 7500,
"train_samples_per_second": 0.52,
"train_steps_per_second": 0.004
}