DUAL-GPO
/

phi-2-gpo-test-longest-iter-v1-3

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

phi-2-gpo-test-longest-iter-v1-3 / train_results.json

BraylonDash's picture

Model save

f36a8d3 verified 3 months ago

raw history blame contribute delete

No virus

197 Bytes

	{
	"epoch": 3.97,
	"train_loss": 0.0014411601897013643,
	"train_runtime": 2755.6406,
	"train_samples": 61135,
	"train_samples_per_second": 1.452,
	"train_steps_per_second": 0.09
	}