Results 121 to 130 of about 379,828 (307)
Hardware-Friendly Static Quantization Method for Video Diffusion Transformers [PDF]
arXivDiffusion Transformers for video generation have gained significant research interest since the impressive performance of SORA. Efficient deployment of such generative-AI models on GPUs has been demonstrated with dynamic quantization. However, resource-constrained devices cannot support dynamic quantization, and need static quantization of the models ...
arxiv
Ambiguities in Quantized Field Theories [PDF]
, 1951 Hiroshi Fukuda, T. Kinoshita
openalex +1 more source
Medical Image Hybrid Watermark Algorithm Based on Frequency Domain Processing and Inception v3
Advanced Intelligent Systems, EarlyView.Frequency domain processing and Inceptionv3 method is used in this article for digital image watermarking of medical images. The algorithm extracts feature vectors from medical images using discrete wavelet transform and discrete cosine transform, enhancing the robustness of the image watermark With the widespread use of digital medical images in ...
Yu Fan +4 more
wiley +1 more source
Pioneering 4-Bit FP Quantization for Diffusion Models: Mixup-Sign Quantization and Timestep-Aware Fine-Tuning [PDF]
arXivModel quantization reduces the bit-width of weights and activations, improving memory efficiency and inference speed in diffusion models. However, achieving 4-bit quantization remains challenging. Existing methods, primarily based on integer quantization and post-training quantization fine-tuning, struggle with inconsistent performance. Inspired by the
arxiv
Binary Quantization of Signal Amplitudes: Effect for Radar Angular Accuracy [PDF]
, 1964 Duane H. Cooper
openalex +1 more source
Multi‐Diseases Detection with Memristive System on Chip
Advanced Intelligent Systems, EarlyView.A robust disease detection system, which is capable of the early prevention of acute myocardial infarction and the detection of liver cancer, is implemented on a memristive system‐on‐chip (SoC). A fully integrated SoC is utilized to ensure the system's portability, low latency, high accuracy, and energy efficiency for medical analysis.
Zihan Wang +7 more
wiley +1 more source
CrossQuant: A Post-Training Quantization Method with Smaller Quantization Kernel for Precise Large Language Model Compression [PDF]
arXivPost-Training Quantization (PTQ) is an effective technique for compressing Large Language Models (LLMs). While many studies focus on quantizing both weights and activations, it is still a challenge to maintain the accuracy of LLM after activating quantization.
arxiv
The Bohr-Sommerfeld quantization rule and the Weyl correspondence [PDF]
, 1965 Petros N. Argyres
openalex +1 more source
Advanced Intelligent Systems, EarlyView.To argue the applicability of continuum wall‐climbing robots in the nuclear industry, an electromagnetic adhering and self‐inductive sensing integrated claw is developed. The flexible claw realizes an integrated design of the physical structure and excitation circuit. An automatic contact strategy based on fuzzy controller is established, which enables
Pengpeng Yang +3 more
wiley +1 more source
Angewandte Chemie, EarlyView.This Minireview highlights recent advances in understanding the role of lattice dynamics and electron–phonon coupling in organic semiconductors. Emphasis is placed on mode‐resolved insights into dynamics‐induced electronic disorder, emerging strategies for phonon engineering, and the implications of both low‐ and high‐frequency vibrations on charge ...
Barbara M. T. C. Peluzo +4 more
wiley +2 more sources