https://github.com/AbhiLegend/OpenVinoQuantization

https://github.com/AbhiLegend/OpenVinoQuantization

Last synced: about 1 month ago
JSON representation

• Host: GitHub
• URL: https://github.com/AbhiLegend/OpenVinoQuantization
• Owner: AbhiLegend
• Created: 2023-10-23T09:23:39.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2023-10-23T09:29:14.000Z (about 1 year ago)
• Last Synced: 2024-08-01T21:47:38.023Z (4 months ago)
• Language: Jupyter Notebook
• Size: 4.88 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-openvino - OpenVINO Quantization - Image Quantization Classification using STL 10 Dataset. (Table of content / Miscellaneous)

README

        
# OpenVinoQuantization

Of course! Let's summarize the steps in the provided code:

1. **Setup**: The necessary libraries are imported. The possible labels for the STL10 dataset are defined, and the data directory is set.

2. **Data Preparation**:

   - Data transformations, which convert images to tensors and normalize them using a previously computed mean and standard deviation, are set up.

   - The STL10 dataset's 'test' split is loaded using these transformations, and a DataLoader is created.

3. **Image and Label Collection**: All images and labels from the DataLoader are collected into separate lists for easy access.

4. **Visualization & Inference**:

   - A function is created to plot images given their indices.

   - Another function performs inference on given images using a specified model and returns predicted labels.

   - Four specific images are selected and displayed.

   - Inference is run on these images using two different models (a float model and a quantized model), and the results are printed out.

In summary, the code prepares and visualizes a subset of the STL10 dataset, and then it demonstrates inference on this subset using two different models, showing the predicted labels for each.