alibaba/BladeDISC
Alibaba's dynamic shape compiler for optimizing machine learning inference workloads across PyTorch and TensorFlow.
Not currently ranked — collecting fresh signals.
star history
BladeDISC is an end-to-end compiler for machine learning workloads that focuses on dynamic shape optimization. It supports frontends including PyTorch 2.0 and TensorFlow, and builds on MLIR for compiler infrastructure. The project targets both inference and training workloads, providing compilation optimization across CPUs and GPUs with deployment solutions for production environments.
Frequently asked
- What is alibaba/BladeDISC?
- Alibaba's dynamic shape compiler for optimizing machine learning inference workloads across PyTorch and TensorFlow.
- Is BladeDISC open source?
- Yes — alibaba/BladeDISC is open source, released under the Apache-2.0 license.
- What language is BladeDISC written in?
- alibaba/BladeDISC is primarily written in C++.
- How popular is BladeDISC?
- alibaba/BladeDISC has 932 stars on GitHub.
- Where can I find BladeDISC?
- alibaba/BladeDISC is on GitHub at https://github.com/alibaba/BladeDISC.