← Quantization
Papers with Code · CC BY-SA 4.0

ImageNet

아카이브 평가 27건 · 마지막 적재/변경

평가 기록

과거 수집 기록이며 최신 순위가 아닙니다. 데이터셋 버전·split·실험 조건과 지표 방향이 검증되지 않은 결과는 직접 비교할 수 없습니다. 날짜는 논문 날짜이며 최신 평가일을 뜻하지 않습니다.

표가 화면보다 넓으면 좌우로 스크롤하여 논문과 구현 링크를 확인하세요.

원본 기록 순서 · 순위 미부여
모델원본 지표·값논문구현
FQ-ViT (ViT-L)
Top-1 Accuracy (%)
85.03
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
FQ-ViT (ViT-B)
Top-1 Accuracy (%)
83.31
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
FQ-ViT (Swin-B)
Top-1 Accuracy (%)
82.97
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
FQ-ViT (Swin-S)
Top-1 Accuracy (%)
82.71
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
FQ-ViT (DeiT-B)
Top-1 Accuracy (%)
81.20
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
FQ-ViT (Swin-T)
Top-1 Accuracy (%)
80.51
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
FQ-ViT (DeiT-S)
Top-1 Accuracy (%)
79.17
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
Xception W8A8
Top-1 Accuracy (%)
78.972
Weight bits
8
Activation bits
8
HPTQ: Hardware-Friendly Post Training Quantization
2021-09-19
코드 1
ADLIK-MO-ResNet50-W4A4
Top-1 Accuracy (%)
77.878
Weight bits
4
Activation bits
4
Learned Step Size Quantization
2019-02-21
코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
ADLIK-MO-ResNet50-W3A4
Top-1 Accuracy (%)
77.34
Weight bits
3
Activation bits
4
Learned Step Size Quantization
2019-02-21
코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
EfficientNet-B0 ReLU W8A8
Top-1 Accuracy (%)
77.092
Weight bits
8
Activation bits
8
HPTQ: Hardware-Friendly Post Training Quantization
2021-09-19
코드 1
ResNet50-W4A4 (paper)
Top-1 Accuracy (%)
76.7
Weight bits
4
Activation bits
4
Learned Step Size Quantization
2019-02-21
코드 1 · 코드 2 · 코드 3 · 코드 4 · 코드 5 · 코드 6 · 코드 7 · 코드 8 · 코드 9
EfficientNet-B0-W8A8
Top-1 Accuracy (%)
76.4
Weight bits
8
Activation bits
8
HMQ: Hardware Friendly Mixed Precision Quantization Block for CNNs
2020-07-20
코드 1 · 코드 2
EfficientNet-B0-W4A4
Top-1 Accuracy (%)
76
Weight bits
4
Activation bits
4
HMQ: Hardware Friendly Mixed Precision Quantization Block for CNNs
2020-07-20
코드 1 · 코드 2
ResNet50-W3A4
Top-1 Accuracy (%)
75.45
Weight bits
3
Activation bits
4
HMQ: Hardware Friendly Mixed Precision Quantization Block for CNNs
2020-07-20
코드 1 · 코드 2
EfficientNet-B0 W8A8
Top-1 Accuracy (%)
74.216
Weight bits
8
Activation bits
8
HPTQ: Hardware-Friendly Post Training Quantization
2021-09-19
코드 1
MPT (80) +BN
Top-1 Accuracy (%)
74.03
Multi-Prize Lottery Ticket Hypothesis: Finding Accurate Binary Neural Networks by Pruning A Randomly Weighted Network
2021-03-17
코드 1
EfficientNet-W4A4
Top-1 Accuracy (%)
73.8
Weight bits
4
Activation bits
4
LSQ+: Improving low-bit quantization through learnable offsets and better initialization
2020-04-20
코드 1 · 코드 2 · 코드 3 · 코드 4
DenseNet-121 W8A8
Top-1 Accuracy (%)
73.356
Weight bits
8
Activation bits
8
HPTQ: Hardware-Friendly Post Training Quantization
2021-09-19
코드 1
MixNet-W4A4
Top-1 Accuracy (%)
71.7
Weight bits
4
Activation bits
4
LSQ+: Improving low-bit quantization through learnable offsets and better initialization
2020-04-20
코드 1 · 코드 2 · 코드 3 · 코드 4
FQ-ViT (DeiT-T)
Top-1 Accuracy (%)
71.61
Weight bits
8
Activation bits
8
FQ-ViT: Post-Training Quantization for Fully Quantized Vision Transformer
2021-11-27
코드 1
UniQ (Ours)
Top-1 Accuracy (%)
71.5
Training Multi-bit Quantized and Binarized Networks with A Learnable Symmetric Quantizer
2021-04-01
코드 1
MobileNetV2 W8A8
Top-1 Accuracy (%)
71.46
Weight bits
8
Activation bits
8
HPTQ: Hardware-Friendly Post Training Quantization
2021-09-19
코드 1
MobileNetV2
Top-1 Accuracy (%)
70.9
HMQ: Hardware Friendly Mixed Precision Quantization Block for CNNs
2020-07-20
코드 1 · 코드 2
MobileNet-v1 + EWGS + R2Loss
Top-1 Accuracy (%)
69.79
Weight bits
4
R2 Loss: Range Restriction Loss for Model Compression and Quantization
2023-03-14
연결 없음
MobileNet-v1 + LSQ + R2Loss
Top-1 Accuracy (%)
69.64
R2 Loss: Range Restriction Loss for Model Compression and Quantization
2023-03-14
연결 없음
ResNet-18 + PACT + R2Loss
Top-1 Accuracy (%)
68.45
Weight bits
2
Activation bits
4
R2 Loss: Range Restriction Loss for Model Compression and Quantization
2023-03-14
연결 없음