Edit on GitHub

Live.log_image()

Saves the given image val to the output file name.

def log_image(name: str, val):

Usage

from dvclive import Live

with Live(cache_images=True) as live:
    # 1. Log an image from a numpy array:
    import numpy as np
    img_numpy = np.ones((500, 500), np.uint8) * 255
    live.log_image("numpy.png", img_numpy)

    # 2. Or log a matplotlib figure:
    from matplotlib import pyplot as plt
    fig, ax = plt.subplots()
    ax.plot([1, 2, 3, 4])
    live.log_image("matplotlib.png", fig)

    # 3. Or log a `PIL.image`:
    from PIL import Image
    img_pil = Image.new("RGB", (500, 500), (250, 250, 250))
    live.log_image("pil.png", img_pil)

    # 4. Or log an existing image:
    live.log_image("sample.png", "run/batch_0_sample.png")

Description

Supported values for val are:

  • A valid NumPy array (convertible to an image via PIL.Image.fromarray)
  • A matplotlib.figure.Figure instance
  • A PIL.Image instance
  • A path to an image file (str or Path). It should be in a format that is readable by PIL.Image.open()

The images will be saved in {Live.plots_dir}/images/{name}. When using Live(cache_images=True), the images directory will also be cached as part of Live.end(). In that case, a .dvc file will be saved to track it, and the directory will be added to a .gitignore file to prevent Git tracking:

dvclive
โ””โ”€โ”€ plots
    โ”œโ”€โ”€ .gitignore
    โ”œโ”€โ”€ images
    โ”‚   โ”œโ”€โ”€ numpy.png
    โ”‚   โ”œโ”€โ”€ matplotlib.png
    โ”‚   โ”œโ”€โ”€ pil.png
    โ”‚   โ””โ”€โ”€ sample.png
    โ””โ”€โ”€ images.dvc

The logged images can be visualized with dvc plots:

$ dvc plots diff dvclive/plots

Images per step

By default the images will be overwritten on each step. However, you can log images using the following pattern live.log_image(f"folder/{live.step}.png", img):

import numpy as np
from dvclive import Live

with Live() as live:
    base_img = np.ones((500, 500), np.uint8)
    for i in range(10):
      live.log_image(
        f"numpy/{live.step}.png", base_img * i * 10)
      live.next_step()

In DVC Studio and the DVC Extension for VSCode, folders following this pattern will be rendered using an image slider:

DVCLive Studio Image Slider

DVCLive VSCode Image Slider

Parameters

  • name - name of the image file that this command will output.

  • val - image to be saved. See the list of supported values in the Description.

Exceptions

  • dvclive.error.InvalidDataTypeError - thrown if the provided val does not have a supported type.
Content

๐Ÿ› Found an issue? Let us know! Or fix it:

Edit on GitHub

โ“ Have a question? Join our chat, we will help you:

Discord Chat