Encountered [error 11001] getaddrinfo failed

[PeterL-1111]

The original script defined local url: demo.launch(server_name="0.0.0.0", server_port=80)

but I could not open it, I tried to modify the url to demo.launch(server_name="http://127.0.0.1", server_port=7860) which works on stable diffusion on my machine,

then I get the [error 11001] getaddrinfo failed

Please advise, help me😅

[Yangr116]

Hi, please try this one :demo.launch(server_name="0.0.0.0", server_port=7860).

The port 7860 shouldn't be occupied and closed. You can use lsof -i:7860 to check it. Additionally, the port should be opened. If the port is closed, you should use the sudo ufw allow ${your port} in ubuntu.

[PeterL-1111]

Thank you!!

I changed to that and it worked with a bit modification, instead of opening the 0.0.0.0:7860, I was able get to the gradio using http://localhost:7860/

The new issues is I get stuck on "Entering new AgentExecutor chain" after giving instruction such as "change background" or "Generate an image of a happy vicuna running in the grass"(in the demo example)

I was using the command line provided in the demo python gpt4tools.py --base_model E:\vicuna\LLaMA --lora_model E:\vicuna\GPT4Tools\lora_weight --llm_device "cpu" --load "Text2Box_cuda:0,Segmenting_cuda:0,Inpainting_cuda:0,ImageCaptioning_cuda:0"

[Yangr116]

This may be your cpu memory is not enough. You can change the llama to 8bit during loading following here.

[PeterL-1111]

Thank you for the reply!!

I tried to add the 8bit to True and device map to auto, but I encountered this error here. Could you please advise?

Also, my gradio interface is different from the demo, could you let me know how to get the options in the demo?(1st one is demo, 2nd one is my interface)

[Yangr116]

Can you remove the device_map='auto'?

The gradio interface used in the demo will be released.

[PeterL-1111]

I see, and I tried the fix, but it returned this error instead, it looks like there is some compatibility with the 8bit format