Efficient inference on IMG Series4 NNAs
February 6, 2023
Research into neural network architectures generally prioritises accuracy over efficiency. Certain papers have investigated efficiency (Tan and Le 2020) (Sandler, et al. 2018), but quite often this is with CPU- or GPU-based rather than accelerator-based inference in mind.
In this original work from Imagination’s AI Research team, many well-known classification networks trained on ImageNet are evaluated. We are not interested in accuracy or cost in their own right, but rather in efficiency, which is a combination of the two. In other words, we want networks that get high accuracy on our IMG Series4 NNAs at as low a cost as possible. We cover:
- identifying ImageNet classification network architectures that give the best accuracy/performance trade-offs on our Series4 NNAs.
- reducing cost dramatically using quantisation-aware training (QAT) and low-precision weights without affecting accuracy.
To read the full article, click here
Related Semiconductor IP
- Message filter
- SSL/TLS Offload Engine
- TCP/UDP Offload Engine
- JPEG-LS Encoder IP
- JPEG XS - Low-Latency Video
Related Blogs
- Dialling it up on PowerVR GPUs: how to optimise automotive dashboards for efficient rendering
- Deep learning inference performance on the Yitian 710
- IP Cannot be an Efficient Abstraction Level Without SystemC!
- Highly Efficient Video Coding with ARM
Latest Blogs
- What’s Your Vector? Synopsys Introduces New ARC VPX6 Digital Signal Processor
- Rambus CXL IP: A Journey from Spec to Compliance
- Alphawave Semi is in Play!
- Vision Transformers Have Already Overtaken CNNs: Here’s Why and What’s Needed for Best Performance
- Tailoring Root Of Trust Security Capabilities To Specific Customer Needs