In any servers
Intel Core i5 6th Gen or more
NVIDIA GTX 1050 Ti or more (not necessary)
Operating System : Windows 10/Ubuntu 18.04 LTS
In embedded / edge devices
Arm Cortex A52 (64-bit) or more
Additional GPU (NVIDIA only)
While using a Raspberry Pi we recommend to also connect an Intel Neural Compute Stick 2 with it.
Operating System : Ubuntu 18.04 LTS/Debian 9
1. Create Conda environment and install dependencies.
conda create -n env
source activate env
pip install -r requirements.txt
2. Install Broker [rabbitmq] and start its service.
$ sudo systemctl start rabbitmq-server
3. Create user name, password, set credentials, create vhost accordingly. See CELERY_BROKER_URL in [setting file](locallibrary/setting.py)
# create user
$ sudo rabbitmqctl add_user catalog catalog
# create virtual host
$ sudo rabbitmqctl add_vhost catalog
# set permission of virtual host to user catalog
$ sudo rabbitmqctl set_permissions -p catalog catalog ".*" ".*" ".*"
4. Get sports1m model weight:
- Download file rename and save in 'c3d/trained_models/c3d_sports1m.h5'
5. Open new terminal, activate conda env and run Celery worker server. More information in [celery]
celery -A locallibrary worker -l info
6. Start Django web app:
python manage.py makemigrations
python manage.py migrate
python manage.py collectstatic
python manage.py runserver
7. Open browser to http://127.0.0.1:8000 .
Install Docker Desktop (Applicable for Mac/Windows only)
Do all necessary steps which are displayed in the Installation page in Docker Desktop
Fire up Docker Desktop
Open up Terminal (in Mac) or Command Prompt (in Windows) and type -
docker pull arijitstudent/dtechtive
The Docker Container contains all the necessary applications installed beforehand.
Now just follow the steps given here.
Note :- Don't use the term sudo in the Docker Container as it will return an error as because of not having admin rights.