[GH-ISSUE #7220] Reference failure for user uploaded models in OpenAI compatible API #30342

Closed
opened 2026-04-22 09:54:57 -05:00 by GiteaMirror · 3 comments
Owner

Originally created by @aretrace on GitHub (Oct 16, 2024).
Original GitHub issue: https://github.com/ollama/ollama/issues/7220

What is the issue?

I am attempting to call adrienbrault/nous-hermes2theta-llama3-8b via the OpenAI node SDK, I get a 500 with a model ... not found, try pulling it first error (I already have it and chatted with it).

I did a simple curl test:

curl -X POST http://localhost:11434/api/generate -d '{
  "model": "adrienbrault/nous-hermes2theta-llama3-8b",
  "prompt": "Hello, how are you?"
}'

Same 500 error (other library models are working).

A poke at http://localhost:11434/v1/models shows its there:

{
  "object": "list",
  "data": [
    {
      "id": "adrienbrault/nous-hermes2theta-llama3-8b:q4_K_M",
      "object": "model",
      "created": 1729046834,
      "owned_by": "adrienbrault"
    },
    ...
}

I have tried several logical permutations of the model name but none have worked.

OS

macOS

GPU

Apple

CPU

Apple

Ollama version

0.3.13

Originally created by @aretrace on GitHub (Oct 16, 2024). Original GitHub issue: https://github.com/ollama/ollama/issues/7220 ### What is the issue? I am attempting to call `adrienbrault/nous-hermes2theta-llama3-8b` via the OpenAI node SDK, I get a 500 with a `model ... not found, try pulling it first` error (I already have it and chatted with it). I did a simple curl test: ```zsh curl -X POST http://localhost:11434/api/generate -d '{ "model": "adrienbrault/nous-hermes2theta-llama3-8b", "prompt": "Hello, how are you?" }' ``` Same 500 error (other `library` models are working). A poke at `http://localhost:11434/v1/models` shows its there: ```javascript { "object": "list", "data": [ { "id": "adrienbrault/nous-hermes2theta-llama3-8b:q4_K_M", "object": "model", "created": 1729046834, "owned_by": "adrienbrault" }, ... } ``` I have tried several logical permutations of the model name but none have worked. ### OS macOS ### GPU Apple ### CPU Apple ### Ollama version 0.3.13
GiteaMirror added the bug label 2026-04-22 09:54:57 -05:00
Author
Owner

@0x376h commented on GitHub (Oct 16, 2024):

try use code like this?
curl -X POST http://localhost:11434/api/generate -d '{ "model": "adrienbrault/nous-hermes2theta-llama3-8b:q4_K_M", "prompt": "Hello, how are you?" }'

<!-- gh-comment-id:2415949094 --> @0x376h commented on GitHub (Oct 16, 2024): try use code like this? `curl -X POST http://localhost:11434/api/generate -d '{ "model": "adrienbrault/nous-hermes2theta-llama3-8b:q4_K_M", "prompt": "Hello, how are you?" }'`
Author
Owner

@rick-github commented on GitHub (Oct 16, 2024):

0x376h is correct. If you don't specify the quant, it uses latest, so when it returns a 500 error, it's trying to find the model adrienbrault/nous-hermes2theta-llama3-8b:latest.

<!-- gh-comment-id:2415980620 --> @rick-github commented on GitHub (Oct 16, 2024): 0x376h is correct. If you don't specify the quant, it uses `latest`, so when it returns a 500 error, it's trying to find the model `adrienbrault/nous-hermes2theta-llama3-8b:latest`.
Author
Owner

@aretrace commented on GitHub (Oct 16, 2024):

I could've sworn I tried that!
It works, false alarm 🫠

<!-- gh-comment-id:2418141002 --> @aretrace commented on GitHub (Oct 16, 2024): I could've sworn I tried that! It works, false alarm 🫠
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#30342