Error: "No such model 'openai-compat'" when using VSS Engine with VLM_MODEL_TO_USE=openai-compat

Accelerated Computing Intelligent Video Analytics Visual AI Agent
1

I’m encountering an error when running the vss-engine:2.3.0 container. The error is:

No such model 'openai-compat'

Here’s my setup:
.env configuration:

VIA_IMAGE=nvcr.io/nvidia/blueprint/vss-engine:2.3.0
VLM_MODEL_TO_USE=openai-compat

Request sent via cURL:

curl -X POST http://localhost:8100/summarize \
  -H "Content-Type: application/json" \
  -d '{ 
    ...
    "model": "openai-compat",
    ...
  }'

From the logs and API response, it seems like openai-compat is not recognized as a valid model.

3

It’s a valid model type, you can refer to our source code vlm_pipeline.py.

Could you please elaborate on your requirements and the operation steps?

5

It has to use the name gpt-4o to be correct.

.env configuration:

VIA_IMAGE=nvcr.io/nvidia/blueprint/vss-engine:2.3.0
VLM_MODEL_TO_USE=openai-compat
VIA_VLM_OPENAI_MODEL_DEPLOYMENT_NAME=gpt-4o

curl -X POST http://localhost:8100/summarize \
  -H "Content-Type: application/json" \
  -d '{ 
    ...
    "model": "gpt-4o",
    "api_type": "openai",
    ...
  }'
6

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.