https://github.com/moshfiqur/dockerized-faster-rcnn
A docerized micro-service implementation of object detection using Faster RCNN
https://github.com/moshfiqur/dockerized-faster-rcnn
docker gunicorn object-detection python tensorflow
Last synced: 2 months ago
JSON representation
A docerized micro-service implementation of object detection using Faster RCNN
- Host: GitHub
- URL: https://github.com/moshfiqur/dockerized-faster-rcnn
- Owner: moshfiqur
- Created: 2018-12-25T21:07:51.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2018-12-25T21:24:57.000Z (over 7 years ago)
- Last Synced: 2025-02-25T14:38:37.961Z (over 1 year ago)
- Topics: docker, gunicorn, object-detection, python, tensorflow
- Language: Python
- Size: 5.86 KB
- Stars: 0
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## Object detection using Faster-RCNN in Docker
A dockerized implementation of Faster-RCNN object detection running on gunicorn.
### Prerequisite
- The tensorflow models zoo available here https://github.com/tensorflow/models. The downloaded path should be used instead of `/srv/downloads/ml-datasets/tf-models`.
- The already trained Faster-RCNN model named faster_rcnn_inception_resnet_v2_atrous_lowproposals_oid_2018_01_28
can be downloaded from [here][1]. The downloaded models should be available as in `/srv/downloads/ml-datastes/pretrained-models/faster_rcnn_resnet_v2`.
[1]: http://download.tensorflow.org/models/object_detection/faster_rcnn_inception_resnet_v2_atrous_lowproposals_coco_2018_01_28.tar.gz
For other interesting trained models, check [here][2].
[2]: https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md
### Build command
```
$ docker build -t faster-rcnn .
```
### Run command
Run with interactive shell access to container
```
$ docker run -it \
-v /srv/downloads/ml-datasets/tf-models:/usr/src/app/tf-models:ro \
-v /srv/downloads/ml-datasets/pretrained-models/faster_rcnn_resnet_v2:/usr/src/app/models:ro \
faster-rcnn bash
```
Run to test the app in action. We needed to increase the memory limit using `-m 4g` for the object detection model to work.
```
$ docker run \
-v /srv/downloads/ml-datasets/tf-models:/usr/src/app/tf-models:ro \
-v /srv/downloads/ml-datasets/pretrained-models/faster_rcnn_resnet_v2:/usr/src/app/models:ro \
-p 10080:10080 \
-m 4g faster-rcnn
```