Extensible application to help with file system tasks such as test summarization and image captioning.
- Project migrated to .NET 6.
- Image-captioner model served using Flask.
Deploy server locally in Docker container:
$ docker build --pull --rm -f "Dockerfile" -t tensorflow/tensorflow:icmserver-gpu "."
$ docker run --rm -it --network=host --gpus all \
--mount type=bind,source=/path/to/data/train2014,target=/tmp/icmserver/data/train2014,readonly \
tensorflow/tensorflow:icmserver-gpu
Training data should be located on host if using --mount/bind.
Serve model from container:
$ python icmserver.py --serve
Generate captions from the FSH ImageCaptioner plugin.