arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2406.06237cs.LG
收藏

基于推理时解码的高效神经压缩

Efficient Neural Compression with Inference-time Decoding

CEA List(法国原子能和替代能源委员会 List 研究所);Université Paris-Saclay(巴黎萨克雷大学);CEA Leti(法国原子能和替代能源委员会 Leti 研究所)

C. Metz, O. Bichler, A. Dupret

AI总结:

本文提出混合精度、零点量化与熵编码相结合的方法,突破1比特前沿限制,在ImageNet上以低于1%的精度损失实现ResNet的高效压缩,并设计了低延迟的推理兼容解码器。

摘要:

本文探讨了神经网络量化与熵编码相结合以最小化内存占用的问题。量化模型的边缘部署受到精度与位宽权衡的严苛帕累托前沿的制约,导致在低于某一位宽时精度急剧下降。借助混合精度量化,可以缓解这种精度损失,从而实现更灵活的位宽分配。然而,标准混合精度的收益仍然有限,原因在于1比特前沿迫使每个参数至少用1比特数据进行编码。本文提出一种结合混合精度、零点量化和熵编码的方法,将ResNet的压缩边界推进到1比特前沿之外,在ImageNet基准上精度下降低于1%。从实现角度看,紧凑的解码器架构具有降低的延迟,从而支持与推理兼容的解码。

英文摘要:

This paper explores the combination of neural network quantization and entropy coding for memory footprint minimization. Edge deployment of quantized models is hampered by the harsh Pareto frontier of the accuracy-to-bitwidth tradeoff, causing dramatic accuracy loss below a certain bitwidth. This accuracy loss can be alleviated thanks to mixed precision quantization, allowing for more flexible bitwidth allocation. However, standard mixed precision benefits remain limited due to the 1-bit frontier, that forces each parameter to be encoded on at least 1 bit of data. This paper introduces an approach that combines mixed precision, zero-point quantization and entropy coding to push the compression boundary of Resnets beyond the 1-bit frontier with an accuracy drop below 1% on the ImageNet benchmark. From an implementation standpoint, a compact decoder architecture features reduced latency, thus allowing for inference-compatible decoding.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑