baricevic
/

flax70

Text Generation

Model card Files Files and versions

flax70 / params.json

baricevic's picture

End of training

17111c9 verified over 1 year ago

history blame contribute delete

176 Bytes

	{
	"dim": 8192,
	"ffn_dim_multiplier": 1.3,
	"multiple_of": 4096,
	"n_heads": 64,
	"n_kv_heads": 8,
	"n_layers": 80,
	"norm_eps": 1e-05,
	"vocab_size": 128256,
	"rope_theta": 500000.0
	}