Papers with Code 아카이브
Quantization
태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.
- 하위 태스크
- 2
- 평가표
- 10
- 보존된 평가 기록
- 32
Quantization is a promising technique to reduce the computation cost of neural network training, which can replace high-cost floating-point numbers (e.g., float32) with low-cost fixed-point numbers (e.g., int8/int16).
Source: Adaptive Precision Training: Quantify Back Propagation in Neural Networks with Fixed-point Numbers
코드
관련 논문과 평가 결과에 연결된 구현입니다. 출처에서 확인된 저자 구현에는 공식 코드 표시가 있습니다.
https://github.com/zihaoliu123/DeepN-Jpeg
Feature Distillation: DNN-Oriented JPEG Compression Against Adversarial Examples
https://github.com/zihaoliu123/Feature-Distillation-DNN-Oriented-JPEG-Compression-Against-Adversarial-Examples
Feature Distillation: DNN-Oriented JPEG Compression Against Adversarial Examples
https://github.com/zirui-ray-liu/exact
ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training
https://github.com/ziye2chen/LLMs-for-Mathematical-Analysis
https://github.com/zju-vipa/mosaickd
Data-Free Network Quantization With Adversarial Knowledge Distillation
https://github.com/zomux/neuralcompressor
Compressing Word Embeddings via Deep Compositional Code Learning
https://github.com/zzzxxxttt/Pytorch_DoReFaNet
DoReFa-Net: Training Low Bitwidth Convolutional Neural Networks with Low Bitwidth Gradients
출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.