ondevicellm
/

phi_1_5_dpo_ep6

Text Generation

alignment-handbook

Generated from Trainer

text-generation-inference

Inference Endpoints

Model card Files Files and versions Metrics Training metrics Community

phi_1_5_dpo_ep6 / train_results.json

hushell's picture

init (#1)

f70168c verified 10 months ago

194 Bytes

	{
	"epoch": 6.0,
	"train_loss": 0.4762616708060343,
	"train_runtime": 91379.3933,
	"train_samples": 58917,
	"train_samples_per_second": 3.869,
	"train_steps_per_second": 0.06
	}