.png){kind=icon}
返回
Loss aware post-training quantization
.png){kind=icon}
OA
AI
DOI：10.1007/s10994-021-06053-z.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Neural network quantization enables the deployment of large models on resource-constrained devices. Current post-training quantization methods fall short in terms of accuracy for INT4 (or lower) but provide reasonable accuracy for INT8 (or above). In this work, we study the effect of quantization on the structure of the loss landscape. We show that the structure is flat and separable for mild quantization, enabling straightforward post-training quantization methods to achieve good results. We show that with more aggressive quantization, the loss landscape becomes highly non-separable with steep curvature, making the selection of quantization parameters more challenging. Armed with this understanding, we design a method that quantizes the layer parameters jointly, enabling significant accuracy improvement over current post-training quantization methods. Reference implementation is available at lattps://github.conilynahshaninn-quantization-pytorch/tree/masterilapq.
Keyword:
Convolutional neural networks
Post-training quantization
期刊
IF:
2.9
论文数:
2.7K
被引数:
3.4W
机构
引用论文
Numerical Analysis of Thermal Stress for a Stack of Planar Solid Oxide Fuel Cells平板固体氧化物燃料电池堆热应力的数值分析.png)
EnergiesOAAI
EnergiesIF0
OAAI分享收藏
查看更多内容