kvquant 0.0.1

More for Keys, Less for Values: Adaptive KV Cache Quantization 🐍🚀🎉🦕

More for Keys, Less for Values: Adaptive KV Cache Quantization ☝️🔑👇🔢

Installation

To install the package, use pip:

pip install kvq

Usage

To use the package, import it in your Python code:

import kvq

medviz.layered_plot(image_path="dataset/1-1.nii", mask_paths=["dataset/small_bowel.nii", "dataset/1-1-label.nii"], mask_colors=["red", "yellow"], title="Layered Plot")

The layered_plot function creates a layered plot of an image and one or more masks. The masks are overlaid on top of the image using the specified colors. The resulting plot can be used to visualize the location of structures or regions of interest in the image.

import medviz

medviz.gif(
    image_path="dataset/1-1.nii",
    mask_paths=[
        "dataset/small_bowel.nii",
        "dataset/1-1-label.nii",
        "dataset/vertebrae_L3.nii.gz",
        "dataset/vertebrae_L4.nii.gz",
        "dataset/vertebrae_L5.nii.gz",
    ],
    mask_colors=["red", "yellow", "green", "blue", "purple"],
    title="Expert Annotations",
    interval=70,
    start_slice=30,
    end_slice=130,
    save_path="animation.gif",
)

The gif function creates an animated GIF of an image and one or more masks. The masks are overlaid on top of the image using the specified colors. The resulting GIF can be used to visualize the location of structures or regions of interest in the image.

GitHub repository: https://github.com/mohsenhariri/kvq