Prior works on Post-training Quantization (PTQ) typically separate a neural network into sub-nets and quantize them sequentially. This process pays little attention to the dependency across the sub-nets, hence is less optimal. In this paper, we propose a novel Network-Wise Quantization (NWQ) approach to fully leveraging inter-layer dependency. NWQ faces a larger scale combinatorial optimization problem of discrete variables than in previous works, which raises two major challenges: over-fitting and discrete optimization problem. NWQ alleviates over fitting via a Activation Regularization (AR) technique, which better controls the activation distribution. To optimize discrete variables, NWQ introduces Annealing Softmax (ASoftmax) and Annealing Mixup (AMixup) to progressively transition quantized weights and activations from continuity to discretization, respectively. Extensive experiments demonstrates that NWQ outperforms prior state-of-the-art approaches by a large margin: 20.24% for the challenging configuration of MobileNetV2 with 2 bits on ImageNet, pushing extremely low-bit PTQ from feasibility to usability. In addition, NWQ is able to achieve competitive or better results with only 10% computation cost of previous works.

This paper focuses on PTQ.

Post-training quantization (PTQ) has stood out as a cost-effective and promising model compression paradigm in recent years, as it avoids computationally intensive model retraining. Nevertheless, current PTQ methods for Vision Transformers (ViTs) still suffer from significant accuracy degradation, especially under low-bit quantization. To address these shortcomings, we analyze the prevailing Hessian-guided quantization loss, and uncover certain limitations of conventional Hessian approximations. By following the block-wise reconstruction framework, we propose a novel PTQ method for ViTs, dubbed FIMA-Q. Specifically, we firstly establish the connection between KL divergence and FIM, which enables fast computation of the quantization loss during reconstruction. We further propose an efficient FIM approximation method, namely DPLR-FIM, by employing the diagonal plus low-rank principle, and formulate the ultimate quantization loss. Our extensive experiments, conducted across various vision tasks with representative ViT-based architectures on public datasets, demonstrate that our method substantially promotes the accuracy compared to the state-of-the-art approaches, especially in the case of low-bit quantization. The source code is available at https://github.com/ShiheWang/FIMA-Q.

Post-training model quantization is a widely adopted technique for reducing the memory and computational costs of large language models (LLMs). However, most existing methods rely on uniform or heuristic bitwidth assignments, failing to account for the nonuniform sensitivity of weights to quantization noise. In this paper, we propose a novel framework for allocating quantization bitwidths based on sensitivity metrics derived from a Hessian proxy. We make key assumptions, which allow the layer/component-wise loss function to be expressed as an explicit function of the bitwidths. This enables a neat formulation of the bit allocation problem as a convex optimization task, whose closed-form solution adapts precision across weights to minimize the layer-wise quantization loss. Inspecting the solution provides several insights (such as the equal-loss structure), which are then exploited to design the proposed \textbf{BAQ} (Bit Allocation Quantization) algorithm. The proposed algorithm achieves a good trade-off between loss minimization and complexity and allows BAQ to be integrated into standard quantization pipelines with minimal overhead. Experimental results show that BAQ consistently outperforms GPTQ, achieving up to 56$\times$ lower perplexity at the same bitwidth on large language models ranging from 125M to 30B parameters. Leveraging our analytical results derived from solving the optimal bit allocation problem, we also provide a theoretical explanation for the observed gains. All codes of this paper are available at https://github.com/CSU-ModelCompression/BAQ.

Large Language Models (LLMs) have demonstrated remarkable success across a wide range of language tasks, but their deployment on edge devices remains challenging due to the substantial memory requirements imposed by their large parameter sizes. Weight-only quantization presents a promising solution to reduce the memory footprint of LLMs. However, existing approaches primarily focus on integer-bit quantization, limiting their adaptability to fractional-bit quantization tasks and preventing the full utilization of available storage space on devices. In this paper, we introduce Channel-Wise Mixed-Precision Quantization (CMPQ), a novel mixed-precision quantization method that allocates quantization precision in a channel-wise pattern based on activation distributions. By assigning different precision levels to different weight channels, CMPQ can adapt to any bit-width constraint. CMPQ employs a non-uniform quantization strategy and incorporates two outlier extraction techniques that collaboratively preserve the critical information, thereby minimizing the quantization loss. Experiments on different sizes of LLMs demonstrate that CMPQ not only enhances performance in integer-bit quantization tasks but also achieves significant performance gains with a modest increase in memory usage. CMPQ thus represents an adaptive and effective approach to LLM quantization, offering substantial benefits across diverse device capabilities. Large Language Models (LLMs), trained on massive text corpora and containing up to hundreds of billions of parameters, have demonstrated exceptional performance across a wide range of language tasks (Brown, 2020; Chowdhery et al., 2023; Du et al., 2022; Hoffmann et al., 2022; Thoppilan et al., 2022; Touvron et al., 2023a).

The dimensionality of the embedding and the number of available embeddings ( also called codebook size) are critical factors influencing the performance of Vector Quantization(VQ), a discretization process used in many models such as the Vector Quantized Variational Autoencoder (VQ-VAE) architecture. This study examines the balance between the codebook sizes and dimensions of embeddings in VQ, while maintaining their product constant. Traditionally, these hyper parameters are static during training; however, our findings indicate that augmenting the codebook size while simultaneously reducing the embedding dimension can significantly boost the effectiveness of the VQ-VAE. As a result, the strategic selection of codebook size and embedding dimensions, while preserving the capacity of the discrete codebook space, is critically important. To address this, we propose a novel adaptive dynamic quantization approach, underpinned by the Gumbel-Softmax mechanism, which allows the model to autonomously determine the optimal codebook configuration for each data instance. This dynamic discretizer gives the VQ-VAE remarkable flexibility. Thorough empirical evaluations across multiple benchmark datasets validate the notable performance enhancements achieved by our approach, highlighting the significant potential of adaptive dynamic quantization to improve model performance.

Recent studies introduced effective compression techniques for Large Language Models (LLMs) via post-training quantization or low-bit weight representation. Although quantized weights offer storage efficiency and allow for faster inference, existing works have indicated that quantization might compromise performance and exacerbate biases in LLMs. This study investigates the confidence and calibration of quantized models, considering factors such as language model type and scale as contributors to quantization loss. Firstly, we reveal that quantization with GPTQ to 4-bit results in a decrease in confidence regarding true labels, with varying impacts observed among different language models. Secondly, we observe fluctuations in the impact on confidence across different scales. Finally, we propose an explanation for quantization loss based on confidence levels, indicating that quantization disproportionately affects samples where the full model exhibited low confidence levels in the first place.

Large language models (LLMs) have shown remarkable capabilities in various tasks. However their huge model size and the consequent demand for computational and memory resources also pose challenges to model deployment. Currently, 4-bit post-training quantization (PTQ) has achieved some success in LLMs, reducing the memory footprint by approximately 75% compared to FP16 models, albeit with some accuracy loss. In this paper, we propose SmoothQuant+, an accurate and efficient 4-bit weight-only PTQ that requires no additional training, which enables lossless in accuracy for LLMs for the first time. Based on the fact that the loss of weight quantization is amplified by the activation outliers, SmoothQuant+ smoothes the activation outliers by channel before quantization, while adjusting the corresponding weights for mathematical equivalence, and then performs group-wise 4-bit weight quantization for linear layers. We have integrated SmoothQuant+ into the vLLM framework, an advanced high-throughput inference engine specially developed for LLMs, and equipped it with an efficient W4A16 CUDA kernels, so that vLLM can seamlessly support SmoothQuant+ 4-bit weight quantization. Our results show that, with SmoothQuant+, the Code Llama-34B model can be quantized and deployed on a A100 40GB GPU, achieving lossless accuracy and a throughput increase of 1.9 to 4.0 times compared to the FP16 model deployed on two A100 40GB GPUs. Moreover, the latency per token is only 68% of the FP16 model deployed on two A100 40GB GPUs. This is the state-of-the-art 4-bit weight quantization for LLMs as we know.

Neural networks are getting better accuracy with higher energy and computational cost. After quantization, the cost can be greatly saved, and the quantized models are more hardware friendly with acceptable accuracy loss. On the other hand, recent research has found that neural networks are vulnerable to adversarial attacks, and the robustness of a neural network model can only be improved with defense methods, such as adversarial training. In this work, we find that adversarially-trained neural networks are more vulnerable to quantization loss than plain models. To minimize both the adversarial and the quantization losses simultaneously and to make the quantized model robust, we propose a layer-wise adversarial-aware quantization method, using the Lipschitz constant to choose the best quantization parameter settings for a neural network. We theoretically derive the losses and prove the consistency of our metric selection. The experiment results show that our method can effectively and efficiently improve the robustness of quantized adversarially-trained neural networks.

The Massive Multiple Input Multiple Output (MIMO) system is a core technology of the next generation communication. With the growing complexity of CSI in massive MIMO system, traditional compressive sensing based CSI feedback has become a bottleneck problem that is limited in piratical. Recently, numerous deep learning based CSI feedback approaches demonstrate the efficiency and potential. However, the existing methods lack a reasonable interpretation of the deep learning model and the accuracy of the model decreases significantly as the CSI compression rate increases. In this paper, from the intrinsic properties of CSI data itself, we devised the corresponding deep learning building blocks to compose a novel neural network CQNet and experiment result shows CQNet outperform the state-of-the-art method with less computational overhead by achieving an average performance improvement of 8.07% in both outdoor and indoor scenarios. In addition, this paper also investigates the reasons for the decrease in model accuracy at large compression rates and proposes a strategy to embed a quantization layer to achieve effective compression, by which the original accuracy loss of 67.19% on average is reduced to 21.96% on average, and the compression rate is increased by 8 times on the original benchmark. The massive multiple-input multiple-output (MIMO) technology is considered one of the core technologies of the next generation communication system, e.g., 5G. By equipping large number of antennas, base station (BS) can sufficiently utilize spatial diversity to improve channel capacity.