HuggingFaceTB/SmolLM2-135M-Instruct
4241.7m
1{2 "epoch": 1.9973828840617638,3 "eval_logits/chosen": 4.600930213928223,4 "eval_logits/rejected": 4.9520039558410645,5 "eval_logps/chosen": -443.648193359375,6 "eval_logps/rejected": -378.15826416015625,7 "eval_loss": 0.6740825176239014,8 "eval_rewards/accuracies": 0.6150793433189392,9 "eval_rewards/chosen": -0.07192634046077728,10 "eval_rewards/margins": 0.26874542236328125,11 "eval_rewards/rejected": -0.34067174792289734,12 "eval_runtime": 20.4479,13 "eval_samples": 2000,14 "eval_samples_per_second": 97.81,15 "eval_steps_per_second": 3.08116}