meta_chat_reasoning_25_75_system / train_results.json
ryanmarten's picture
End of training
49d0624 verified
{
"epoch": 5.0,
"total_flos": 2.0080798451032392e+18,
"train_loss": 0.36193929355636784,
"train_runtime": 24487.1768,
"train_samples_per_second": 6.452,
"train_steps_per_second": 0.05
}