carlyou
/

DeepSeek-Coder-V2-Lite-Instruct-NVFP4

8-bit precision

compressed-tensors

Model card Files Files and versions

DeepSeek-Coder-V2-Lite-Instruct-NVFP4 / generation_config.json

carlyou's picture

Add DeepSeek-Coder-V2-Lite-Instruct-NVFP4 quantized to NVFP4 via llm-compressor

4537259 verified 5 months ago

History Blame Contribute Delete

181 Bytes

	{
	"_from_model_config": true,
	"bos_token_id": 100000,
	"do_sample": true,
	"eos_token_id": 100001,
	"temperature": 0.3,
	"top_p": 0.95,
	"transformers_version": "4.57.6"
	}