client can not connect the server

[duoduo07]

Error: Unable to look up model info. Are you sure you're running an activation server for this dataset? Current URL: http://localhost:8002

server and client are deployed on the same machine,
server start as the command:
python neuron_explainer/activation_server/main.py --run_model False --port 8000

client start as the command:
NEURON_VIEWER_ACTIVATION_SERVER_URL=http://localhost:8000 npm start

Is the OPENAI_API_KEY neccssary?

[starlitsky2010]

Hi OpenAI guys,

I also met this problem:

Step 1:
transformer-debugger$ export export OPENAI_API_KEY=sk-xxxxx
transformer-debugger$ python neuron_explainer/activation_server/main.py --model_name gpt2-small --port 8005
Step 2:
transformer-debugger/neuron_viewer$ NEURON_VIEWER_ACTIVATION_SERVER_URL=http://10.0.80.43:8005 npm start

[email protected] start
parcel src/index.html --no-cache

Server running at http://localhost:1234
✨ Built in 9.41s

Visit http://10.0.80.43:1234/, it prints:

Neuron viewer
Loading...

Visit http://10.0.80.43:1234/gpt2-small/tdb_alpha

An error occurs:
Error: Unable to look up model info. Are you sure you're running an activation server for this dataset? Current URL: http://10.0.80.43:8005

Thanks

[starlitsky2010]

I

Error: Unable to look up model info. Are you sure you're running an activation server for this dataset? Current URL: http://localhost:8002

server and client are deployed on the same machine, server start as the command: python neuron_explainer/activation_server/main.py --run_model False --port 8000

client start as the command: NEURON_VIEWER_ACTIVATION_SERVER_URL=http://localhost:8000 npm start

Is the OPENAI_API_KEY neccssary?

Hi duoduo07,

Have you fixed this problem?

Thanks

[XinyuZhou2000]

hi, same problem.
in neuron_viewer/src/requests/paths.ts try this:

export function getActivationServerUrl(): string {
  if (process.env.NEURON_VIEWER_ACTIVATION_SERVER_URL) {
    return process.env.NEURON_VIEWER_ACTIVATION_SERVER_URL;
  }
  return "http://localhost:8000";
}

 #change 0.0.0.0 into localhost

maybe you're using a school/company's internet

[Sfc-001]

i can not get gpt2-small

[Sfc-001]

the programer will happen : Ran out of input.somebady reslove it