[Ollama] ERROR: Non-retryable error occurred: 404 page not found

[mdzidic]

Description:

I'm running ComfyUI inside a Docker container along with an Ollama local server. I have edited urls.json, and the models load correctly. However, when I attempt to generate an enhanced prompt, I receive the following error from the "Troubleshooting" section:

➤ Begin Log for: Advanced Prompt Enhancer, Node #6:
✦ INFO: Additional parameters input: []
✦ INFO: Setting client to OpenAI Open Source LLM object
✦ INFO: Maximum tries set to: 3
✦ ERROR: Non-retryable error occurred: 404 page not found
✦ ERROR: Request failed: 404 page not found

➤ Begin Log for: Ollama Unload Model Setting:
✦ INFO: URL was validated and is being presented as: http://host.docker.internal:11434/api/generate
✦ INFO: Attempting to set model TTL using URL: http://host.docker.internal:11434/api/generate
✦ INFO: Model unload setting successful.  Response: {"model":"llama3.1:latest","created_at":"2025-02-04T13:17:00.393244595Z","response":"","done":true,"done_reason":"load"}

Expected Behavior:

The enhanced prompt should generate successfully without errors.

Additional Information:

• ComfyUI version: 0.3.13
• Ollama version: 0.3.9

[glibsonoran]

What's the url for your docker setup? http://host.docker.internal:11434/API/generate? The 404 is indicating that it's found the server, but can't find the endpoint (path).

…

On Tue, Feb 4, 2025, 6:59 AM Marko Dzidic ***@***.***> wrote: Description: I'm running ComfyUI inside a Docker container along with an Ollama local server. I have edited urls.json, and the models load correctly. However, when I attempt to generate an enhanced prompt, I receive the following error from the "Troubleshooting" section: ➤ Begin Log for: Advanced Prompt Enhancer, Node #6: ✦ INFO: Additional parameters input: [] ✦ INFO: Setting client to OpenAI Open Source LLM object ✦ INFO: Maximum tries set to: 3 ✦ ERROR: Non-retryable error occurred: 404 page not found ✦ ERROR: Request failed: 404 page not found ➤ Begin Log for: Ollama Unload Model Setting: ✦ INFO: URL was validated and is being presented as: http://host.docker.internal:11434/api/generate ✦ INFO: Attempting to set model TTL using URL: http://host.docker.internal:11434/api/generate ✦ INFO: Model unload setting successful. Response: {"model":"llama3.1:latest","created_at":"2025-02-04T13:17:00.393244595Z","response":"","done":true,"done_reason":"load"} Expected Behavior: The enhanced prompt should generate successfully without errors. Additional Information: - ComfyUI version: 0.3.13 - Ollama version: 0.3.9 — Reply to this email directly, view it on GitHub <#182>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHONJOJABKHR5VWV5F6NUPD2ODBVRAVCNFSM6AAAAABWOUSTS2VHI2DSMVQWIX3LMV43ASLTON2WKOZSHAZTAMZVGY4DSMY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[glibsonoran]

You can also try running it with the http.../generate url using the AI_service: Direct Web Connection (URL)

Ollama: http://host.internal:11434/v1
Direct Web: http://host.internal:11434/v1/chat/completions

[mdzidic]

Ollama URL .../v1 did a trick, thanks!

[glibsonoran]

Good news, enjoy! :)