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
- Mesochronous Bridge for PCIe/CXL
- AI-native GPU
- OpenGMSL Verification IP
- OpenGMSL Leaf IP
- FlexGen Multi-Die Smart Network-on-Chip (NoC) IP
Related Blogs
- Deep learning inference performance on the Yitian 710
- ChiPy®: Bridge Neural Networks and C++ on Silicon — Full Inference Pipelines with Zero CPU Round-Trips
- Introducing Cortex-A35: ARM's Most Efficient Application Processor
- Energy Efficient Designs with MIPI
Latest Blogs
- Automated, Faster Specification to Sign-Off with IDS-AI
- NovaTech Automation Crius PIU: Bringing Conventional Instrument Transformers onto the IEC 61850 Process Bus
- Single Pair Ethernet and TSN: The In-Robot Network Behind the Next Humanoid Robots
- A design path to success exists for ultra-low-voltage SoCs
- Where Routine Flow Ends Veriest Formal Expertise Begins