The installation process for CaImAn GUI is slightly modified from the original CaImAn program. For more details, refer to the official CaImAn documentation.

Adding CaImAn Data Folder

1. Navigate to the folder containing the caiman_gui source folder.
2. Create a new folder named caiman_data.
3. Inside caiman_data, create two folders:
   • temp
   • model
4. Place the CaImAn model files (from the original source code) into the model folder.
5. The folder structure should look like:

user/
├── caiman_gui/
├── caiman_data/ 
│   ├── temp/
│   └── model/
│       ├── cnn_model.h5
│       ├── cnn_model.h5.pb
│       ├── cnn_model.json
│       ├── cnn_model_online.h5
│       └── ...

Cloning the CaImAn GUI program

1. Click on VS Code.
2. Go to Terminal → New Terminal or press CTRL + SHIFT + ‘ (Windows/Linux) or CMD + SHIFT + ‘ (Mac).
3. In the terminal, run:

git clone <https://github.com/kantarcilab/caiman_gui.git>
cd caiman_gui
mamba env create -f environemnt.yml -n caiman
conda activate caiman
pip install -e . 

Installing TensorFlow

The program requires TensorFlow 2.15.0 and Keras 2.15.0 for compatibility with the original CaImAn models (Keras 3 is not supported).

In the activated caiman environment, run: