Quantization

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

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 871–880 of 4925 papers

Title	Date	Tasks	Status	Hype
A holistic approach to polyphonic music transcription with neural networks	Oct 26, 2019	Beat TrackingMusic Transcription	CodeCode Available	1
Q8BERT: Quantized 8Bit BERT	Oct 14, 2019	Linguistic AcceptabilityNatural Language Inference	CodeCode Available	1
Improvements to Target-Based 3D LiDAR to Camera Calibration	Oct 7, 2019	Camera CalibrationPose Estimation	CodeCode Available	1
And the Bit Goes Down: Revisiting the Quantization of Neural Networks	Jul 12, 2019	CPUQuantization	CodeCode Available	1
Data-Free Quantization Through Weight Equalization and Bias Correction	Jun 11, 2019	Data Free Quantizationobject-detection	CodeCode Available	1
Mixed Precision DNNs: All you need is a good parametrization	May 27, 2019	AllQuantization	CodeCode Available	1
Learning to Groove with Inverse Sequence Transformations	May 14, 2019	Generative Adversarial NetworkQuantization	CodeCode Available	1
HAWQ: Hessian AWare Quantization of Neural Networks with Mixed-Precision	Apr 29, 2019	Quantization	CodeCode Available	1
Single Path One-Shot Neural Architecture Search with Uniform Sampling	Mar 31, 2019	Neural Architecture SearchQuantization	CodeCode Available	1
Learned Step Size Quantization	Feb 21, 2019	Model CompressionQuantization	CodeCode Available	1

Show:10 25 50

← PrevPage 88 of 493Next →

All datasets ImageNet CIFAR-10 Wiki-40B AgeDB-30 CFP-FP COCO (Common Objects in Context)IJB-B IJB-C Knowledge-based:LFW

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	FQ-ViT (ViT-L)	Top-1 Accuracy (%)	85.03	—	Unverified
2	FQ-ViT (ViT-B)	Top-1 Accuracy (%)	83.31	—	Unverified
3	FQ-ViT (Swin-B)	Top-1 Accuracy (%)	82.97	—	Unverified
4	FQ-ViT (Swin-S)	Top-1 Accuracy (%)	82.71	—	Unverified
5	FQ-ViT (DeiT-B)	Top-1 Accuracy (%)	81.2	—	Unverified
6	FQ-ViT (Swin-T)	Top-1 Accuracy (%)	80.51	—	Unverified
7	FQ-ViT (DeiT-S)	Top-1 Accuracy (%)	79.17	—	Unverified
8	Xception W8A8	Top-1 Accuracy (%)	78.97	—	Unverified
9	ADLIK-MO-ResNet50-W4A4	Top-1 Accuracy (%)	77.88	—	Unverified
10	ADLIK-MO-ResNet50-W3A4	Top-1 Accuracy (%)	77.34	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	3DCNN_VIVA_3	MAP	160,327.04	—	Unverified
2	DTQ	MAP	0.79	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	OutEffHop-Bert_base	Perplexity	6.3	—	Unverified
2	OutEffHop-Bert_base	Perplexity	6.21	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1		Accuracy	98.13	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1		Accuracy	92.92	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	SSD ResNet50 V1 FPN 640x640	MAP	34.3	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1		TAR @ FAR=1e-4	95.13	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1		TAR @ FAR=1e-4	96.38	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	3DCNN_VIVA_5	All	84,809,664	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1		Accuracy	99.8	—	Unverified