privategpt trying to access openaipublic.blob.core.windows.net

probably i missed something in the docs.

i'm using privategpt since a few weeks without problems installed on debian testing. when i tried to start privategpt to ask a question i got this error:

requests.exceptions.ConnectionError: HTTPSConnectionPool(host='openaipublic.blob.core.windows.net', port=443): Max retries exceeded with url: /gpt-2/encodings/main/vocab.bpe (Caused by NewConnectionError('<urllib3.connection.HTTPSConnection object at 0x7fa2fe5104d0>: Failed to establish a new connection: [Errno -3] Temporary failure in name resolution'))

what's about that connection? in the docs i read it would only connect to the web while installing and never again if i don't change certain things in the config file.

would be great if somebody could explain.

@fawkeyes looks like your firewall or network is blocking out going 443 or dns not resolving

@imartinez i also see the same issue after i blocked outgoing port 443 similar issue and solution here #https://github.com/openai/whisper/discussions/1399#discussioncomment-6155951 i also have this issue open for outgoing connection to AWS #https://github.com/imartinez/privateGPT/issues/1527#issue-2092898628 can you please advise which location to put these files in ? or if there is a way to package with software

Pinging blob.dsm10prdstf02a.store.core.windows.net [20.60.179.33] with 32 bytes of data:
Reply from 20.60.179.33: Destination host unreachable.
Reply from 20.60.179.33: Destination host unreachable.

requests.exceptions.ConnectTimeout: HTTPSConnectionPool(host='openaipublic.blob.core.windows.net', port=443): Max retries exceeded with url: /gpt-2/encodings/main/vocab.bpe (Caused by ConnectTimeoutError(<urllib3.connection.HTTPSConnection object at 0x7f20adbd1450>, 'Connection to openaipublic.blob.core.windows.net timed out. (connect timeout=None)'))

Hit this error today when trying running the docker images with an internal network. I downloading the files and placing them in tiktoken_cache but it didn't work.

I tried updating the volume to include the cache file which downloaded a file looking similar to encoder.json and stopped the error, but the UI (http://localhost:8001/) failed to load.

  private-gpt-ollama:
    volumes:
      - ./local_data:/home/worker/app/local_data
      - ./tiktoken_cache:/home/worker/app/tiktoken_cache