lole25
/

phi-2-gpo-ultrafeedback-lora

alignment-handbook

Generated from Trainer

Model card Files Files and versions Metrics Training metrics Community

phi-2-gpo-ultrafeedback-lora / train_results.json

lole25's picture

Model save

be46998 verified 8 months ago

198 Bytes

	{
	"epoch": 2.0,
	"train_loss": 2.2877212975025803e-05,
	"train_runtime": 862.5384,
	"train_samples": 30567,
	"train_samples_per_second": 70.877,
	"train_steps_per_second": 1.106
	}