ondevicellm
/

tinyllama_moe_sft_ultrachat200k_v2_epochs3

Text Generation

alignment-handbook

Generated from Trainer

Inference Endpoints

text-generation-inference

Model card Files Files and versions Metrics Training metrics Community

tinyllama_moe_sft_ultrachat200k_v2_epochs3 / train_results.json

hushell's picture

Model save

935514f verified 5 months ago

raw history blame contribute delete

No virus

196 Bytes

	{
	"epoch": 3.0,
	"train_loss": 1.1024342694896863,
	"train_runtime": 51497.7783,
	"train_samples": 207865,
	"train_samples_per_second": 8.508,
	"train_steps_per_second": 0.066
	}