English | 日本語

Examples of StreamDiffusion.

If you want to maximize performance, you need to install with following steps explained in README.md in the root directory, and use --acceleration tensorrt option at the end of each command. At default, StreamDiffusion uses xformers for acceleration, which is not the fastest option.

Take a screen capture and process it. This script only works on Windows.

When you run the script, a translucent window appears. Position it at where you want to capture the screen and press the enter key to finalize the capture area.

You need to install extra dependencies for this script as follows:

pip install -r screen/requirements.txt

python screen/main.py

Just measure the performance of StreamDiffusion.

benchmark/multi.py spawns multiple processes for postprocessing and benchmark/single.py does not.

python benchmark/multi.py

With TensorRT acceleration:

python benchmark/multi.py --acceleration tensorrt

examples からサンプルを実行できます。

Using SD-Turbo and TensorRT, perform text-to-image with optimal performance.

optimal-performance/multi.py is optimized for RTX4090 and performs batch processing, while optimal-performance/single.py does not.

python optimal-performance/multi.py

python optimal-performance/single.py

Perform image-to-image.

img2img/multi.py takes a directory of input images and outputs to another directory as arguments, and img2img/single.py takes a single image.

Image-to-image for a single image:

python img2img/single.py --input path/to/input.png --output path/to/output.png

Image-to-image for multiple images:

python img2img/multi.py --input ./input --output-dir ./output

Perform text-to-image.

txt2img/multi.py generates multiple images from a single prompt, and txt2img/single.py generates a single image.

Text-to-image for a single image:

python txt2img/single.py --output output.png --prompt "A cat with a hat"

Text-to-image for multiple images:

python txt2img/multi.py --output ./output --prompt "A cat with a hat"

Perform video-to-video conversion.

You need to install extra dependencies for this script as follows:

pip install -r vid2vid/requirements.txt

python vid2vid/main.py --input path/to/input.mp4 --output path/to/output.mp4