MaskRCNN not serving when using PyTorchServe Docker

Steps which led me to the bug

1. I downloaded PyTorch Serve Docker image from here --> https://hub.docker.com/r/pytorch/torchserve
2. Downloaded https://download.pytorch.org/models/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth to my computer in folder C:/examples
3. Downloaded https://github.com/pytorch/serve/tree/master/examples/ to my computer in folder C:/examples
4. Created empty folder called model-store here --> C:/model-store
5. Run the command docker run -it -p 8080:8080 -p 8081:8081 --name mar -v C:/model-store:/home/model-server/model-store -v C:/examples:/home/model-server/examples pytorch/torchserve:latest
6. Run the command docker exec -it <container_name> /bin/bash
7. From inside the container bash torch-model-archiver --force --model-name maskrcnn --version 1.0 --model-file /home/model-server/examples/object_detector/maskrcnn/model.py --serialized-file /home/model-server/examples/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth --handler object_detector --export-path /home/model-server/model-store --extra-files /home/model-server/examples/object_detector/index_to_name.json
8. From inside the container bash torchserve --start --model-store /home/model-server/model-store --models maskrcnn=maskrcnn.mar
9. Download persons.jpg to my desktop from here --> https://github.com/pytorch/serve/tree/master/examples/object_detector/maskrcnn
10. From my desktop (host machine) curl http://127.0.0.1:8080/predictions/maskrcnn -T persons.jpg
11. I am getting the following error { "code": 404, "type": "ModelNotFoundException", "message": "Model not found: maskrcnn" }

Note :- I followed the same steps for Densenet image classifier (from the examples from the repo) and it worked perfectly fine. MaskRCNN doesn’t work however

[dhanainme]

A good place to debug would be to do a list model to confirm that the model has been registered.

curl -X GET http://127.0.0.1:8081/models should tell you if the model is registered.

[nagamanikandank]

@alvarobartt I had to redownload the image and it worked !!!

Attaching all the steps that worked for me. Thanks a lot !!!

docker pull ubuntu:18.04
docker run -it -p 9080:8080 -p 9081:8081 -p 9082:8082 -p 9070:7070 -p 9071:7071 --name maskrcnn_pytorch ubuntu:18.04
cd home/
apt-get update
apt-get install python3 python3-dev python3-pip openjdk-11-jre-headless git wget curl -y
python3 -m pip install torch --no-cache-dir
python3 -m pip install torchvision --no-cache-dir
python3 -m pip install torch-model-archiver --no-cache-dir
python3 -m pip install torchserve==0.2.0 --no-cache-dir
pip3 show torch
pip3 show torchvision
pip3 show torchserve
git clone https://github.com/pytorch/serve
wget https://download.pytorch.org/models/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth
torch-model-archiver --model-name maskrcnn --version 1.0 --model-file /home/serve/examples/object_detector/maskrcnn/model.py --serialized-file /home/maskrcnn_resnet50_fpn_coco-bf2d0c1e.pth --handler object_detector --extra-files /home/serve/examples/object_detector/index_to_name.json
mkdir model_store
mv maskrcnn.mar model_store/
python3 -m pip install opencv-python pillow requests --no-cache-dir
torchserve --start --model-store /home/model_store --models maskrcnn=maskrcnn.mar

===================================
In a New Docker cli of maskrcnn_pytorch container :-
===================================
curl -X GET http://127.0.0.1:8081/models
cd /home/serve/examples/object_detector
curl -X POST http://127.0.0.1:8080/predictions/maskrcnn -T persons.jpg