Siddharth Kumar Sah b4b59a7500 fix: force CPU mode in download_models.py for build-time compatibility ...

paddlepaddle-gpu tries to load libcuda.so.1 on import, but no GPU
driver exists during Docker build. Set PADDLE_DEVICE=cpu, FLAGS_use_cuda=0,
and CUDA_VISIBLE_DEVICES="" before any ML imports to force CPU mode.

2026-04-10 11:22:56 +08:00

..

.dockerignore

feat: add multi-stage Docker build with Python ML dependencies

2026-03-22 03:09:14 +08:00

docker-compose.test.yml

chore: remove dead code, add test infrastructure, update docs

2026-03-23 11:46:45 +08:00

docker-compose.yml

fix: address code review findings before merge

2026-04-10 10:35:13 +08:00

Dockerfile

fix: split paddlepaddle-gpu and paddleocr installs, use --extra-index-url

2026-04-10 11:01:58 +08:00

Dockerfile.test

chore: remove dead code, add test infrastructure, update docs

2026-03-23 11:46:45 +08:00

download_models.py

fix: force CPU mode in download_models.py for build-time compatibility

2026-04-10 11:22:56 +08:00

entrypoint.sh

fix: log volume permission errors instead of swallowing them

2026-04-04 14:06:48 +08:00