DALI and dali_backend
DALI is a GPU-accelerated data processing library, while the DALI backend is a Triton plugin that integrates DALI pipelines into Triton's inference server, making them **complements** designed to be used together for optimized preprocessing during model serving.
About DALI
NVIDIA/DALI
A GPU-accelerated library containing highly optimized building blocks and an execution engine for data processing to accelerate deep learning training and inference applications.
This tool helps deep learning engineers accelerate their training and inference workflows for models that process image, video, or audio data. It takes raw multimedia files and quickly transforms them with operations like decoding, cropping, and resizing, outputting pre-processed data ready for model ingestion. The primary users are machine learning practitioners developing or deploying deep learning applications who need to optimize data pipelines.
About dali_backend
triton-inference-server/dali_backend
The Triton backend that allows running GPU-accelerated data pre-processing pipelines implemented in DALI's python API.
This tool helps machine learning engineers accelerate the data preparation stage for deep learning models, especially during inference. It takes raw input data, like images or sensor readings, and efficiently processes it using GPU-accelerated pipelines. The output is pre-processed data ready for your deep learning model, speeding up the overall application performance.
Scores updated daily from GitHub, PyPI, and npm data. How scores work