ComfyUI Extension: ComfyUI-Transceiver📡
Run ComfyUI workflows without the setup
No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.
Transceiver is a python library that swiftly exchanges fundamental data structures, specifically numpy arrays, between processes, optimizing AI inference tasks that utilize ComfyUI.
Looking for a different extension?
Custom Nodes (2)
README
ComfyUI-Transceiver📡
Transceiver is a python library that swiftly exchanges fundamental data structures, specifically numpy arrays, between processes, optimizing AI inference tasks that utilize ComfyUI.
Why?
When processing a large number of requests, the SaveImage and LoadImage nodes may be IO-limited, and using shared memory improves performance by passing images only through memory access, not through IO.
Install as ComfyUI custom nodes
cd /path/to/ComfyUI
source venv/bin/activate
cd custom_nodes
git clone https://github.com/nat-chan/comfyui-transceiver
pip install -r requirements.txt
cd ../.. # cd /path/to/ComfyUI
python main.py # launch
Custom Nodes
Save Image Transceiver📡
save image to shared memory
inputs
- image: ImageTensor
- name: str
required name is shared memor identical name
Load Image Transceiver📡
load image from shared memory
inputs
- name: str
required name is shared memor identical name
Run ComfyUI workflows without the setup
No installs, no CUDA version roulette, no GPU sitting idle on your bill. Bring a workflow and run it in the browser.