Qwen
/

Qwen2-VL-72B-Instruct-AWQ

Image-Text-to-Text

4-bit precision

Model card Files Files and versions Community

Qwen2-VL-72B-Instruct-AWQ / generation_config.json

可亲

fix(pad zero) pad intermediate_size to 29696 to make sure quantized model can use 8 tensor-parallel in vllm

688e28d about 2 months ago

history blame contribute delete

227 Bytes

	{
	"chat_format": "chatml",
	"do_sample": true,
	"eos_token_id": 151643,
	"max_new_tokens": 512,
	"max_window_size": 6144,
	"pad_token_id": 151643,
	"top_k": 0,
	"top_p": 0.01,
	"transformers_version": "4.45.0.dev0"
	}