pytc-client

Docker installation instructions

# Pull the image (tag 0.0.1 -- subject to change in future versions)
docker pull xbalbinus/pytc-client:0.0.1
# Expose ports to run the backend servers on Docker
docker run -it -p 4242:4242 -p 4243:4243 -p 4244:4244 -p 6006:6006 --shm-size=8g xbalbinus/pytc-client:0.0.1

Installation

0. Create a Virtual Environment via. Conda

conda create -n pytc python=3.9
conda activate pytc
conda install pytorch torchvision cudatoolkit=11.3 -c pytorch

Alternatively, dependencies can be installed with native Python via. the following:

# Create a venv
python -m venv .venv
# if running in windows, replace the line above with '.\.venv\Scripts\activate.bat'
source .venv/bin/activate

# Install dependencies
pip install torch torchvision cuda-python

In the rare event that your device does not support CUDA, you may run the following respectively:

# If using a conda environment
conda create -n pytc python=3.9
conda activate pytc
conda install pytorch torchvision

# If installing via native python
python -m venv .venv
source .venv/bin/activate 
# if running in windows, replace the line above with '.\.venv\Scripts\activate.bat'
pip install torch torchvision

1. Client

cd client
npm install
npm run build

2. API Server:

cd server_api
pip install -r requirements.txt

3. Pytc-connectomics:

In root folder,

git clone https://github.com/zudi-lin/pytorch_connectomics.git
cd pytorch_connectomics
pip install --editable .

Run Project

To run

# if running on mac or linux:
./start.sh

# if running on windows:
./start.bat

In a separate terminal

cd client
npm run electron

Below is a link to a video demo: showing how to set up and run the app: video demo