bigalex95/pix2pix_TecoGAN

combined Pix2Pix (Conditional GAN) and TecoGAN (TEmporally COherent GAN) super resolution Networks.

Combined Pix2Pix (Conditional GAN) and TecoGAN (TEmporally COherent GAN) super resolution Networks.

Step 1. Clone repository

git clone https://github.com/bigalex95/pix2pix_TecoGAN

Step 2. Create virtualenv and install necessary libraries.

# create virtualenv
python -m venv venv_name
# change environment to virtualenv
## Windows 
### 1. cmd.exe
./venv_name\Scripts\activate
### 2. PowerShell
./venv_name/Scripts/activate.ps1
## Linux bash
source ./venv_name/bin/activate
# install packages
pip install -r requirements.txt

Usage

Step 1. Run jupyter notebook.

jupyter notebook

You should see the notebook open in your browser.

Step 2. Open scripts in pix2pix folder and train pix2pix model.

Step 3. Open scripts in TecoGAN_PyTorch fodler and train tecogan model or download pretrained models.

Step 4. Open scripts in TF_to_TF-TRT folder convert pix2pix checkpoint model to saved_model.

Step 5. Open pix2pix_tecogan.ipynb script, configure necessary arguments and run the script.

Compile to execution file

Step 1. Use auto-py-to-exe library for compiling.

# Windows
auto-py-to-exe.exe
# Linux
auto-py-to-exe

This command open auto-py-to-exe GUI like below.