Minbyul
/

biomistral-7b-dpo-full-sft-wo-medication_qa

Text Generation

alignment-handbook

Generated from Trainer

Inference Endpoints

text-generation-inference

Model card Files Files and versions Community

biomistral-7b-dpo-full-sft-wo-medication_qa / eval_results.json

Minbyul's picture

End of training

0b3f71d verified about 2 months ago

raw history blame contribute delete

No virus

553 Bytes

	{
	"epoch": 1.0,
	"eval_logits/chosen": -4.123082160949707,
	"eval_logits/rejected": -4.0820841789245605,
	"eval_logps/chosen": -456.7123107910156,
	"eval_logps/rejected": -693.0343627929688,
	"eval_loss": 0.4655996561050415,
	"eval_rewards/accuracies": 0.75,
	"eval_rewards/chosen": -0.5065802931785583,
	"eval_rewards/margins": 0.8316203951835632,
	"eval_rewards/rejected": -1.338200569152832,
	"eval_runtime": 12.1695,
	"eval_samples": 76,
	"eval_samples_per_second": 6.245,
	"eval_steps_per_second": 0.247
	}