lewtun
/

gemma-7b-dpo-full-orca-v0

Text Generation

alignment-handbook

Generated from Trainer

Inference Endpoints

text-generation-inference

Model card Files Files and versions Metrics Training metrics Community

gemma-7b-dpo-full-orca-v0 / eval_results.json

lewtun's picture

lewtun HF staff

End of training

0063657 verified 4 months ago

raw history blame contribute delete

No virus

559 Bytes

	{
	"epoch": 0.99,
	"eval_logits/chosen": 94.10542297363281,
	"eval_logits/rejected": 141.18112182617188,
	"eval_logps/chosen": -725.1587524414062,
	"eval_logps/rejected": -860.7156982421875,
	"eval_loss": 0.013092889450490475,
	"eval_rewards/accuracies": 0.9921875,
	"eval_rewards/chosen": 4.552516937255859,
	"eval_rewards/margins": 12.267425537109375,
	"eval_rewards/rejected": -7.714908599853516,
	"eval_runtime": 44.617,
	"eval_samples": 1000,
	"eval_samples_per_second": 22.413,
	"eval_steps_per_second": 0.717
	}