reward_model_train_debug / eval_results.json
shirwu's picture
Training in progress, step 1
bc0885e verified
{
"epoch": 0.0022222222222222222,
"eval_accuracy": 0.45871559633027525,
"eval_loss": 0.693359375,
"eval_runtime": 4.7216,
"eval_samples_per_second": 42.359,
"eval_steps_per_second": 10.59
}