MVQ:Towards Efficient DNN Compression and Acceleration with Masked Vector Quantization

Vector quantization(VQ) is a hardware-friendly DNN compression method that can reduce the storage cost and weight-loading datawidth of hardware accelerators. However, conventional VQ techniques lead to significant accuracy loss because the important weights are not well preserved. To tackle this problem, a novel approach called MVQ is proposed, which aims at better approximating important weights with a limited number of codewords. At the algorithm level, our approach removes the less important weights through N:M pruning and then minimizes the vector clustering error between the remaining weights and codewords by the masked k-means algorithm. Only distances between the unpruned weights and the codewords are computed, which are then used to update the codewords. At the architecture level, our accelerator implements vector quantization on an EWS (Enhanced weight stationary) CNN accelerator and proposes a sparse systolic array design to maximize the benefits brought by masked vector quantization.\\ Our algorithm is validated on various models for image classification, object detection, and segmentation tasks. Experimental results demonstrate that MVQ not only outperforms conventional vector quantization methods at comparable compression ratios but also reduces FLOPs. Under ASIC evaluation, our MVQ accelerator boosts energy efficiency by 2.3$\times$ and reduces the size of the systolic array by 55\% when compared with the base EWS accelerator. Compared to the previous sparse accelerators, MVQ achieves 1.73$\times$ higher energy efficiency.

VQ4ALL: Efficient Neural Network Representation via a Universal Codebook

The rapid growth of the big neural network models puts forward new requirements for lightweight network representation methods. The traditional methods based on model compression have achieved great success, especially VQ technology which realizes the high compression ratio of models by sharing code words. However, because each layer of the network needs to build a code table, the traditional top-down compression technology lacks attention to the underlying commonalities, resulting in limited compression rate and frequent memory access. In this paper, we propose a bottom-up method to share the universal codebook among multiple neural networks, which not only effectively reduces the number of codebooks but also further reduces the memory access and chip area by storing static code tables in the built-in ROM. Specifically, we introduce VQ4ALL, a VQ-based method that utilizes codewords to enable the construction of various neural networks and achieve efficient representations. The core idea of our method is to adopt a kernel density estimation approach to extract a universal codebook and then progressively construct different low-bit networks by updating differentiable assignments. Experimental results demonstrate that VQ4ALL achieves compression rates exceeding 16 $\times$ while preserving high accuracy across multiple network architectures, highlighting its effectiveness and versatility.

Efficiency Meets Fidelity: A Novel Quantization Framework for Stable Diffusion

Text-to-image generation of Stable Diffusion models has achieved notable success due to its remarkable generation ability. However, the repetitive denoising process is computationally intensive during inference, which renders Diffusion models less suitable for real-world applications that require low latency and scalability. Recent studies have employed post-training quantization (PTQ) and quantization-aware training (QAT) methods to compress Diffusion models. Nevertheless, prior research has often neglected to examine the consistency between results generated by quantized models and those from floating-point models. This consistency is crucial in fields such as content creation, design, and edge deployment, as it can significantly enhance both efficiency and system stability for practitioners. To ensure that quantized models generate high-quality and consistent images, we propose an efficient quantization framework for Stable Diffusion models. Our approach features a Serial-to-Parallel calibration pipeline that addresses the consistency of both the calibration and inference processes, as well as ensuring training stability. Based on this pipeline, we further introduce a mix-precision quantization strategy, multi-timestep activation quantization, and time information precalculation techniques to ensure high-fidelity generation in comparison to floating-point models. Through extensive experiments with Stable Diffusion v1-4, v2-1, and XL 1.0, we have demonstrated that our method outperforms the current state-of-the-art techniques when tested on prompts from the COCO validation dataset and the Stable-Diffusion-Prompts dataset. Under W4A8 quantization settings, our approach enhances both distribution similarity and visual similarity by 45%-60%.

VQ4DiT: Efficient Post-Training Vector Quantization for Diffusion Transformers

The Diffusion Transformers Models (DiTs) have transitioned the network architecture from traditional UNets to transformers, demonstrating exceptional capabilities in image generation. Although DiTs have been widely applied to high-definition video generation tasks, their large parameter size hinders inference on edge devices. Vector quantization (VQ) can decompose model weight into a codebook and assignments, allowing extreme weight quantization and significantly reducing memory usage. In this paper, we propose VQ4DiT, a fast post-training vector quantization method for DiTs. We found that traditional VQ methods calibrate only the codebook without calibrating the assignments. This leads to weight sub-vectors being incorrectly assigned to the same assignment, providing inconsistent gradients to the codebook and resulting in a suboptimal result. To address this challenge, VQ4DiT calculates the candidate assignment set for each weight sub-vector based on Euclidean distance and reconstructs the sub-vector based on the weighted average. Then, using the zero-data and block-wise calibration method, the optimal assignment from the set is efficiently selected while calibrating the codebook. VQ4DiT quantizes a DiT XL/2 model on a single NVIDIA A100 GPU within 20 minutes to 5 hours depending on the different quantization settings. Experiments show that VQ4DiT establishes a new state-of-the-art in model size and performance trade-offs, quantizing weights to 2-bit precision while retaining acceptable image generation quality.