Meta-Llama-3-8B-Base-5e-7 / all_results.json
tengxiao1
TX
f952910
{
"epoch": 0.9982631930527722,
"eval_logits/chosen": -0.5905505418777466,
"eval_logits/rejected": -0.5776959657669067,
"eval_logps/chosen": -0.7339656949043274,
"eval_logps/rejected": -0.808925211429596,
"eval_loss": 1.6913511753082275,
"eval_rewards/accuracies": 0.5569105744361877,
"eval_rewards/chosen": -0.7339656949043274,
"eval_rewards/margins": 0.07495953887701035,
"eval_rewards/rejected": -0.808925211429596,
"eval_runtime": 421.6781,
"eval_samples": 1961,
"eval_samples_per_second": 4.65,
"eval_steps_per_second": 0.292,
"total_flos": 0.0,
"train_loss": 1.7001326604879898,
"train_runtime": 20882.5667,
"train_samples": 59876,
"train_samples_per_second": 2.867,
"train_steps_per_second": 0.022
}