| Autor: |
Tokunaga, Hiroyuki, Nicholls, Joel, Vazhenina, Daria, Kanemura, Atsunori |
| Rok vydání: |
2024 |
| Předmět: |
|
| Druh dokumentu: |
Working Paper |
| Popis: |
By quantizing network weights and activations to low bitwidth, we can obtain hardware-friendly and energy-efficient networks. However, existing quantization techniques utilizing the straight-through estimator and piecewise constant functions face the issue of how to represent originally high-bit input data with low-bit values. To fully quantize deep neural networks, we propose pixel embedding, which replaces each float-valued input pixel with a vector of quantized values by using a lookup table. The lookup table or low-bit representation of pixels is differentiable and trainable by backpropagation. Such replacement of inputs with vectors is similar to word embedding in the natural language processing field. Experiments on ImageNet and CIFAR-100 show that pixel embedding reduces the top-5 error gap caused by quantizing the floating points at the first layer to only 1% for the ImageNet dataset, and the top-1 error gap caused by quantizing first and last layers to slightly over 1% for the CIFAR-100 dataset. The usefulness of pixel embedding is further demonstrated by inference time measurements, which demonstrate over 1.7 times speedup compared to floating point precision first layer. |
| Databáze: |
arXiv |
| Externí odkaz: |
|
