nanotron
/

doremi-llama-2.5b-reference

Model card Files Files and versions Community

doremi-llama-2.5b-reference / 70000 /checkpoint_metadata.json

neuralink's picture

add the 70k checkpoint

4a65de7 4 months ago

raw history blame contribute delete

No virus

131 Bytes

	{
	"dp": 8,
	"metas": {
	"consumed_train_samples": 35840000,
	"last_train_step": 70000
	},
	"tp": 8,
	"version": "1.2"
	}