YYYYYYibo
/

nash_dpo_rank4_on_vanilla_iter_1

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

nash_dpo_rank4_on_vanilla_iter_1 / train_results.json

YYYYYYibo's picture

Model save

5ef9bc7 verified 6 months ago

193 Bytes

	{
	"epoch": 1.0,
	"train_loss": 0.6885996239307599,
	"train_runtime": 7326.9096,
	"train_samples": 20000,
	"train_samples_per_second": 2.73,
	"train_steps_per_second": 0.021
	}