Skip to content

Latest commit

 

History

History
52 lines (31 loc) · 1.29 KB

File metadata and controls

52 lines (31 loc) · 1.29 KB

Troubleshooting

Local Qwen2.5-VL inference failed

Real inference requires a CUDA runtime and GPU dependencies. Verify:

python -c "import torch; print(torch.cuda.is_available())"

If this prints False, run only the mocked tests or move to a CUDA environment.

Out of GPU Memory

Reduce the maximum input pixels:

export VISORAG_QWEN_MAX_PIXELS=501760

Use fewer retrieved pages with --top-k, or test on shorter documents.

DOCX Conversion Fails

DOCX conversion requires LibreOffice:

soffice --version

Install LibreOffice and ensure soffice is on PATH.

401 unauthorized

Set and pass the same bearer token:

export VISORAG_API_TOKEN="change-me"
python -m visorag query --token "$VISORAG_API_TOKEN" ...

invalid_top_k

top_k must be an integer from 1 to 20.

file_too_large

The default upload limit is 25 MB. Change VISORAG_MAX_UPLOAD_BYTES only if the runtime has enough CPU, disk, and GPU memory for the larger request.

Import Smoke Fails After Installing GPU Packages

The package is intended to import without loading models. If import allocates CUDA memory or downloads models, inspect recent changes to src/visorag/features/visual_retrieval.py and src/visorag/features/answer_generation.py for eager model loading.