[CS655] Deploy image recognition application on GENI.
- Create a new Slice
- Create new Resource
- Choose RSpec with
File
and uploadgeni_rspec.xml
in this repo. - Click on
Site 1
box to choose one available InstaGENI site. - Click
Reserve Resources
and wait until the resources are finished.
If the final result is Failed
, please select another InstaGENI.
After the ssh login, we need to install and run the server.
sudo apt-get update
sudo apt-get upgrade
sudo apt-get install uwsgi uwsgi-plugin-python
sudo apt install python3-pip
python3 -m pip install --upgrade pip
python3 -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cpu
python3 -m pip install -r server/requirements.txt
uwsgi --plugin http,python3 --http :5000 --wsgi-file server/server.py --callable app
Login in to the client node.
sudo apt-get update
sudo apt install python3-pip
python3 -m pip install --upgrade pip
python3 -m pip install --upgrade Pillow
Make sure the server is on
cd client
wget http://72.36.65.79:5000
Run the client
python3 client.py [-s SERVER] [-p PORT] [-n number of image] [-t timeout] [-f width/height of the image]
- SERVER: http://72.36.65.79
- PORT: 5000
- number of image: 5
- timeout: 1000
- width/height: 100/100
OpenAI CLIP(Contrastive Language-Image Pre-Training)
- Cache mechanism
- TCP handshake HTTP/1.0