YYYYYYibo
/

simple_online_epoch_2_dpo_iter_5

alignment-handbook

Generated from Trainer

Model card Files Files and versions Community

simple_online_epoch_2_dpo_iter_5 / all_results.json

YYYYYYibo's picture

Model save

4264ef2 verified 5 months ago

194 Bytes

	{
	"epoch": 0.99,
	"train_loss": 0.687320882978003,
	"train_runtime": 21824.845,
	"train_samples": 20000,
	"train_samples_per_second": 0.916,
	"train_steps_per_second": 0.007
	}