Frequently asked questions#

Why am I unable to download a model?

There are multiple reasons you might not be able to download a model. Verify that you have a connection to the internet, and check the application logs for HTTP errors. You can view the logs at the following locations:

For Windows machines, you can find the logs here: C:\Users\<USERNAME>\AppData\Roaming\ai-navigator\logs

For macOS machines, you can find the logs here: ~/Library/Logs/ai-navigator/

If you are receiving a 401 error, sign out, then sign back in with your Anaconda Cloud account credentials. If that doesn’t fix your issue, quit and relaunch the application and try again.

Why is my model producing garbage in chat?

The models provided with Anaconda AI Navigator are optimized at various quantization methods to make the model available for users with different hardware. Lower quantization methods have less precise responses and lower quality output than their larger counterparts.

Some models require a large amount of RAM to operate properly. If your system does not have enough resources available, your model may produce “AI hallucinations”, or your system might crash completely. Verify you have enough resources before chatting with a model. If your model does produce a hallucination, delete the prompt that generated the output and try again, or open a new chat and continue working from there.

Error: couldn’t bind HTTP server socket, hostname: 127.0.0.1, port: 8080

If you attempt to load a model into the API server and encounter an error ending with couldn’t bind HTTP server socket, hostname: 127.0.0.1, port: 8080, it indicates that another process is already using the port. To resolve this, identify any running processes associated with the API server and terminate them to free up the port, or try another port.