DUAL-GPO
/

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

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

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

BraylonDash's picture

Model save

5a1daf8 verified 3 months ago

raw history blame contribute delete

No virus

198 Bytes

	{
	"epoch": 3.97,
	"train_loss": 0.0010954971686230911,
	"train_runtime": 8408.6931,
	"train_samples": 61135,
	"train_samples_per_second": 0.476,
	"train_steps_per_second": 0.029
	}