We are thrilled to announce that Adlik, an LF AI & Data Foundation Incubation-Stage Project, released version 0.5.0 on June 21th, called Eagle. Adlik is a toolkit for accelerating deep learning inference, which provides an overall support for bringing trained models into production and eases the learning curves for different kinds of inference frameworks. In Adlik, Model Optimizer and Model Compiler deliver optimized and compiled models for a certain hardware environment, and Inference Engine provides deployment solutions for cloud, edge and device.
In this Eagle version, we can see that Adlik has explored many technologies, e.g., distillation and quantization techniques are utilized in Model Optimizer to optimize the YOLOv5 model, resulting in a performance increase of nearly 2.5 times; Model Compiler now support compilation models from oneflow to onnx; for BERT models, the inference performance was greatly improved after optimization by Adlik, as shown in the benchmark test in Intel 8260 CPU environment; Torch runtime was integrated in Inference Engine engine; A new repository for Model Zoo was created to store the models to be optimized by Adlik and the output, which makes it quite easy for users to quickly exploit the advantage of Adlik.
The detailed features of Adlik Eagle version are as follows:
Model Optimizer
- Support quantization and distillation of YOLOv5s models, which achieves nearly 2.5 times inference performance improvement with OpenVINO runtime.
Model Zoo
- A new repository that stores Adlik optimized and compiled models, including ResNet series models and YOLOv5 series models.
Compiler
- Support compilation path from oneflow to onnx
- OpenVINO upgraded to version 2022.1.0
Inference Engine
- Support Torch runtime
- OpenVINO upgraded to version 2022.1.0
Benchmark Test
- Benchmark test for BERT model on Intel 8260 CPU, including throughput and other performance indicators.
It’s time for all those who want to bring artificial intelligence into production to get on-board, and try Adlik Eagle version. Any feedback is welcomed so we can further improve the project. To learn more about the Adlik 0.5.0 release, check out the full release notes. Want to get involved with Adlik? Be sure to follow our WeChat public account (Adlik_AI) or join the Adlik-Announce and Adlik Technical-Discuss mailing lists to join the community and stay connected on the latest updates.
Adlik Key Links
LF AI Resources
- Learn about membership opportunities
- Explore the interactive landscape
- Check out our technical projects
- Join us at upcoming events
- Read the latest announcements on the blog
- Subscribe to the mailing lists
- Follow us on Twitter or LinkedIn
- Access other resources on LF AI’s GitHub or Wiki
