DUAL-GPO-2
/

phi-2-ipo-renew1

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

phi-2-ipo-renew1 / train_results.json

BraylonDash's picture

Model save

70a5420 verified 2 months ago

raw history blame contribute delete

No virus

195 Bytes

	{
	"epoch": 2.0,
	"train_loss": 2099.8451463309884,
	"train_runtime": 42790.3459,
	"train_samples": 30567,
	"train_samples_per_second": 1.429,
	"train_steps_per_second": 0.089
	}