Minbyul
/

biomistral-7b-dpo-full-wo-live_qa-ep3

Text Generation

alignment-handbook

Generated from Trainer

Inference Endpoints

text-generation-inference

Model card Files Files and versions Community

biomistral-7b-dpo-full-wo-live_qa-ep3 / eval_results.json

Minbyul's picture

End of training

5e19b6b verified 2 months ago

raw history blame contribute delete

No virus

569 Bytes

	{
	"epoch": 1.0,
	"eval_logits/chosen": -1.3924791812896729,
	"eval_logits/rejected": 1.0068244934082031,
	"eval_logps/chosen": -385.703857421875,
	"eval_logps/rejected": -374.45843505859375,
	"eval_loss": 0.5347221493721008,
	"eval_rewards/accuracies": 0.6527777910232544,
	"eval_rewards/chosen": -0.8760086297988892,
	"eval_rewards/margins": 0.9213289618492126,
	"eval_rewards/rejected": -1.7973376512527466,
	"eval_runtime": 61.4466,
	"eval_samples": 568,
	"eval_samples_per_second": 9.244,
	"eval_steps_per_second": 0.293
	}