Approximation-Aware Training for Efficient Neural Network Inference on MRAM Based CiM Architecture
Convolutional neural networks (CNNs), despite their broad applications, are constrained by high computational and memory requirements. Existing compression techniques often neglect approximation errors incurred during training. This work proposes approximation-aware-training, in which group of weigh...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Open Journal of Nanotechnology |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/10819260/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Convolutional neural networks (CNNs), despite their broad applications, are constrained by high computational and memory requirements. Existing compression techniques often neglect approximation errors incurred during training. This work proposes approximation-aware-training, in which group of weights are approximated using a differential approximation function, resulting in a new weight matrix composed of approximation function's coefficients (AFC). The network is trained using backpropagation to minimize the loss function with respect to AFC matrix with linear and quadratic approximation functions preserving accuracy at high compression rates. This work extends to implement an compute-in-memory architecture for inference operations of approximate neural networks. This architecture includes a mapping algorithm that modulates inputs and map AFC to crossbar arrays directly, eliminating the need to predict approximated weights for evaluating output. This reduces the number of crossbars, lowering area and energy consumption. Integrating magnetic random-access memory-based devices further enhances performance by reducing latency and energy consumption. Simulation results on approximated LeNet-5, VGG8, AlexNet, and ResNet18 models trained on the CIFAR-100 dataset showed reductions of 54%, 30%, 67%, and 20% in the total number of crossbars, respectively, resulting in improved area efficiency. In the ResNet18 architecture, latency and energy consumption decreased by 95% and 93.3% with spin-orbit torque (SOT) based crossbars compared to RRAM-based architectures. |
|---|---|
| ISSN: | 2644-1292 |