Console/terminal sample for Class Activation Map (CAM) visualization with an ImageNet-pretrained InceptionV3 in PyTorch, supporting Windows 11 and Ubuntu Linux 24.04. The script classifies an image, prints the top-k ImageNet predictions, and saves a figure with the original image plus one CAM heatmap overlay per class, showing where the network looks. It works out of the box with the bundled example image and weights.

 

[Folder Structure]

  • CLAUDE.md / AGENT.md: program specification (identical content).
  • CHANGELOG.md: change log.
  • copyright.txt: copyright notice.
  • data/cam/: cat_dog.png (the example image) and imagenet_class_index.json (the 1000-class map).
  • model/cam/: inception_v3_google-1a9a5a14.pth (ImageNet weights); the CAM figures ({name}_cam.png) are written here too.
  • src/pretrainedmodels/: the vendored Cadene pretrainedmodels package.
  • tool/tools.py: shared helpers.
  • src/: source code, launcher scripts and this document.
  • venv/: Python virtual environment, created by step 01.

[Instructions]

1. 01_venv_create.bat / 01_venv_create.sh

Delete the old venv/, create a fresh one and install src/requirements.txt. Aborts loudly if pip fails.

 

2. 02_venv_activate.bat / 02_venv_activate.sh

Activate the virtual environment. On Linux use source src/02_venv_activate.sh.

 

3. 03_cam.bat / 03_cam.sh

Classify the image and save the CAM figure to model/cam/.

 

4. 99_venv_clear.bat / 99_venv_clear.sh

Delete the venv/ folder.

 

[Python Entry Point]

src/cam.py

  • --source data/cam/cat_dog.png: path to the input image
  • --top 3: how many classes to report
  • --weights model/cam/inception_v3_google-1a9a5a14.pth: the weights file to use
  • --labels data/cam/imagenet_class_index.json: the class-name map
  • --output model/cam/{name}_cam.png: path of the output figure
  • --show: also open a window showing the result (optional, requires a desktop session)

The script prints the top-k labels with their probabilities; the saved figure stacks the original image and one jet-colormap heatmap overlay per class.

 

[Quick Start]

Windows 11

src\01_venv_create.bat

src\03_cam.bat

 

Ubuntu Linux 24.04

bash src/01_venv_create.sh

bash src/03_cam.sh

 

Point --source at your own image, for example: src\03_cam.bat --source path\to\photo.jpg --top 5

 

[References]

This SDK is built in AppForAI - AI Dev Tools.

 

Purchase license separately:  USD 400, permanent authorization, single APP authorization, single machine authorization, one-year activation, one-year download, one-year update, one-year email technical support.