dball
/

zephyr-tiny-dpo-qlora

alignment-handbook

Generated from Trainer

Model card Files Files and versions Metrics Training metrics Community

zephyr-tiny-dpo-qlora / train_results.json

dball's picture

Model save

eb0e803 verified 9 months ago

196 Bytes

	{
	"epoch": 1.0,
	"train_loss": 0.6145847465156994,
	"train_runtime": 171708.6447,
	"train_samples": 61135,
	"train_samples_per_second": 0.356,
	"train_steps_per_second": 0.044
	}