Check out this demo of Qualcomm Technologies’ AIMET, the AI Model Efficiency Toolkit.
AIMET is a library that provides users with advanced quantization and compression techniques from Qualcomm AI Research for trained neural network models.
In this video, Sam is joined by Abhijit Khobare, director of software engineering at Qualcomm Technologies. The two explore why AIMET is an important tool for developers working with edge and mobile devices, the benefits of quantization, and the use of the toolkit to apply advanced quantization techniques.
Play Video about TWIML Qualcomm Model Quantization and Compression for Edge Devices with AIMET Cover
References
- AIMET (AI Model Efficiency Toolkit)
- Deep Learning is Eating 5G. Here’s How. with Joseph Soriaga - #525
- Full-Stack AI Systems Development with Murali Akula - #563
- Advances in Neural Compression with Auke Wiggers - #570
- Optical Flow Estimation, Panoptic Segmentation, and Vision Transformers with Fatih Porikli - #579
- Equivariant Priors for Compressed Sensing with Arash Behboodi - #584
