DUAL-GPO
/

phi-2-gpo-test-longest-iter-random2-1

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

phi-2-gpo-test-longest-iter-random2-1 / train_results.json

BraylonDash's picture

Model save

6f885d6 verified 3 months ago

raw history blame contribute delete

No virus

196 Bytes

	{
	"epoch": 3.97,
	"train_loss": 0.00108777264728316,
	"train_runtime": 2657.7329,
	"train_samples": 61135,
	"train_samples_per_second": 1.505,
	"train_steps_per_second": 0.093
	}