cognitivecomputations
/

DeepSeek-V3-AWQ

Text Generation

4-bit precision

Model card Files Files and versions Community

DeepSeek-V3-AWQ / generation_config.json

v2ray's picture

Fixed cache and added prefill ability.

604068d 21 days ago

137 Bytes

	{
	"_from_model_config": true,
	"bos_token_id": 0,
	"do_sample": true,
	"eos_token_id": 1,
	"transformers_version": "4.48.0.dev0"
	}