DUAL-GPO
/

phi-2-gpo-v25-i1

alignment-handbook

Generated from Trainer

Model card Files Files and versions Metrics Training metrics Community

phi-2-gpo-v25-i1 / train_results.json

lole25's picture

Model save

25a262a verified 5 months ago

No virus

192 Bytes

	{
	"epoch": 1.0,
	"train_loss": 0.105732237171923,
	"train_runtime": 2052.0868,
	"train_samples": 15000,
	"train_samples_per_second": 7.31,
	"train_steps_per_second": 0.228
	}