generation_config.json · majentik/Mistral-Small-4-119B-RotorQuant-MLX-8bit at main

Mistral-Small-4-119B-RotorQuant-MLX-8bit / generation_config.json

majentik's picture

Add MLX 8-bit quantized model with KV cache compression

21d6a54 verified 3 days ago

history blame contribute delete

132 Bytes

	{
	"bos_token_id": 1,
	"eos_token_id": 2,
	"max_length": 1048576,
	"pad_token_id": 11,
	"transformers_version": "5.3.0.dev0"
	}