DUAL-GPO
/

phi-2-gpo-test-longest-iter-random1-0

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

phi-2-gpo-test-longest-iter-random1-0 / train_results.json

BraylonDash's picture

Model save

c2e0269 verified 3 months ago

raw history blame contribute delete

No virus

197 Bytes

	{
	"epoch": 2.67,
	"train_loss": 0.00021988208027323708,
	"train_runtime": 13.6913,
	"train_samples": 61135,
	"train_samples_per_second": 2.922,
	"train_steps_per_second": 0.292
	}