Benefit of pruning and clustering a neural network for before deploying on Arm Ethos-U NPU
July 24, 2023
Pruning and clustering are optimization techniques:
- Pruning: setting weights to zero
- Clustering: grouping weights together into clusters
These techniques modify the weights of a Machine Learning model. In some cases, they enable:
- Significant speed-up of the inference execution
- Reduction of the memory footprint
- Reduction in the overall power consumption of the system
We assume that you can optimize your workload without loss in accuracy and that you target an Arm® Ethos NPU. You can therefore prune and cluster your neural network before using the Vela compiler and deploying it on the Ethos-U hardware. See below for more information on optimizing your workload.
To read the full article, click here
Related Semiconductor IP
- HBM4 PHY IP
- Ultra-Low-Power LPDDR3/LPDDR2/DDR3L Combo Subsystem
- MIPI D-PHY and FPD-Link (LVDS) Combinational Transmitter for TSMC 22nm ULP
- HBM4 Controller IP
- IPSEC AES-256-GCM (Standalone IPsec)
Related Blogs
- Reviewing different Neural Network Models for Multi-Agent games on Arm using Unity
- Neural Network Model quantization on mobile
- Silicon-proven LVTS for 2nm: a new era of accuracy and integration in thermal monitoring
- Ultra Ethernet Consortium Set to Enable Scaling of Networking Interconnects for AI and HPC
Latest Blogs
- ReRAM in Automotive SoCs: When Every Nanosecond Counts
- AndeSentry – Andes’ Security Platform
- Formally verifying AVX2 rejection sampling for ML-KEM
- Integrating PQC into StrongSwan: ML-KEM integration for IPsec/IKEv2
- Breaking the Bandwidth Barrier: Enabling Celestial AI’s Photonic Fabric™ with Custom ESD IP on TSMC’s 5nm Platform