Quantization-aware training can recover accuracy lost by post-training quantization.
problem Post-training quantization (PTQ) can fail sharply at aggressive bitwidths.
method A unified geometric framework that explains PTQ failure and QAT recovery.
result QAT has a useful bias that steers iterates back into the basin.
FrostNet improves INT8 quantization efficiency in mobile networks.
problem The importance of network architecture for optimal INT8 quantization.
method Quantization-aware training (QAT) with StatAssist and GradBoost, hardware-aware NAS.
result FrostNets achieve higher recognition accuracy with comparable latency when quantized.
We present an overview of techniques for quantizing convolutional neural networks for inference with integer weights and activations. Per-channel quantization of weights and per-layer quantization of activations to 8-bits of precision post-training produces classification accuracies within 2% of floating point networks…
Unified finetuning of all quantization degrees of freedom achieves state-of-the-art 4-bit quantization.
problem Achieving high accuracy in quantized neural networks while maintaining speed and resource constraints.
method Quantization-aware finetuning (QFT) that jointly optimizes all quantization degrees of freedom.
result 4-bit weight quantization results on-par with state-of-the-art (SoTA) within PTQ constraints.
APQ jointly optimizes neural architecture, pruning, and quantization for efficient inference.
problem Efficient deep learning inference on resource-constrained hardware.
method Joint optimization of neural architecture, pruning, and quantization policy using a quantization-aware accuracy predictor.
result Joint optimization leads to 2.3% higher ImageNet accuracy with reduced latency and energy consumption.
Degree-Quant improves GNN efficiency by quantizing them without losing accuracy.
problem Efficiency of graph neural networks at inference time.
method Architecturally-agnostic method, Degree-Quant, for quantizing GNNs.
result Degree-Quant trained models perform as well as full-precision models and achieve up to 26% gains.
We extend quantization-aware training to extreme model compression.
problem Maximizing model accuracy with minimal model size.
method Quantize a random subset of weights during training, allowing unbiased gradients through other weights.
result Established new state-of-the-art compromises between accuracy and model size.
Generative model for HW-aware DNN quantization.
problem Efficiently tuning DNNs for various hardware platforms.
method Generative model AQGAN for generating quantization configurations based on target accuracy.
result Generative model generates quantization configurations with competitive accuracy and lower search cost.
This study improves Ernie's accuracy for INT8 inference by modifying its training process.
problem Improving the accuracy of pre-trained models like Ernie for low precision inference.
method Integrates a regularizer into the training process to make it more robust to quantization.
result Increased INT8 accuracy for Ernie models.
QuantEase optimizes LLMs with CD-based quantization, achieving state-of-the-art performance.
problem Efficiently quantize large language models for deployment.
method Layer-wise quantization using CD-based algorithms with matrix and vector operations.
result State-of-the-art performance in perplexity and zero-shot accuracy.
We analyze the effect of quantizing weights and activations of neural networks on their loss and derive a simple regularization scheme that improves robustness against post-training quantization. By training quantization-ready networks, our approach enables storing a single set of weights that can be quantized on-deman…
StatQAT optimizes quantization for deep networks, reducing computational cost and memory usage.
problem Optimal quantization parameters selection for deep neural networks with diverse data distributions.
method Statistical error analysis framework for uniform and floating-point quantization, iterative and analytic quantizers designed for arbitrary and Gaussian-like distributions.
result Improved accuracy and stability in training low-precision neural networks.
Paper develops a statistical framework for quantized training of deep neural networks.
problem Lack of theoretical understanding of gradient quantization in FQT.
method Presented a statistical framework for analyzing FQT algorithms, viewing quantized gradient as a stochastic estimator of QAT gradient.
result Developed two novel gradient quantizers with smaller variance than existing per-tensor quantizer.
Discretizing multi-dimensional data distributions is a fundamental step of modern indexing methods. State-of-the-art techniques learn parameters of quantizers on training data for optimal performance, thus adapting quantizers to the data. In this work, we propose to reverse this paradigm and adapt the data to the quant…
There is a mismatch between the standard theoretical analyses of statistical machine learning and how learning is used in practice. The foundational assumption supporting the theory is that we can represent features and models using real-valued parameters. In practice, however, we do not use real numbers at any point d…
Paper proposes a method to speed up DNNs by quantizing Winograd/Toom-Cook convolutions.
problem Speeding up convolution computations in DNNs with reduced time consumption and improved accuracy.
method Application of base change technique for quantized Winograd-aware training model.
result 8-bit quantized network achieves nearly the same accuracy as direct quantized convolution with minimal additional operations.
The implementation of Deep Convolutional Neural Networks (ConvNets) on tiny end-nodes with limited non-volatile memory space calls for smart compression strategies capable of shrinking the footprint yet preserving predictive accuracy. There exist a number of strategies for this purpose, from those that play with the to…
Winograd convolutions are used to improve quantized neural networks.
problem Improving quantized neural networks using Winograd convolutions.
method Proposed a Winograd-aware formulation of convolution layers to expose numerical inaccuracies to learning.
result Up to 10% higher classification accuracy on CIFAR-10 with Winograd-aware layers.
The study analyzes convergence of adaptive optimizers under low-precision training.
problem Understanding why low-precision training remains effective for large models.
method Developed a theoretical framework for analyzing convergence of adaptive optimizers under floating-point quantization.
result Adaptive optimizers retain convergence rates close to full-precision methods under logarithmic mantissa scaling.
Low-precision representation of deep neural networks (DNNs) is critical for efficient deployment of deep learning application on embedded platforms, however, converting the network to low precision degrades its performance. Crucially, networks that are designed for embedded applications usually suffer from increased de…
BiTAT improves neural network quantization for edge devices by focusing on weight dependencies and disentangling them.
problem Performance degradation of compact neural networks under extreme quantization.
method Task-dependent Aggregated Transformation (BiTAT) method that orthonormalizes weights and progressively quantizes them.
result BiTAT effectively preserves model performance on ImageNet and CIFAR-100 with compact backbones.
Two methods reduce BN and DNN complexity, balancing size and accuracy.
problem Balancing model size and prediction accuracy in Bayesian networks and deep neural networks.
method Quantization-aware training and tree-augmented naive Bayes structure learning extension.
result Pareto optimal models found for small-scale scenarios.
New method uses geometric properties for better density estimation.
problem Uncertainty quantification in ambiguous tasks.
method Winner-takes-all training with centroidal Voronoi tessellations.
result Improved quantization and density estimation.
SOM-VQ tokenizes discrete models with semantic structure and navigable topology.
problem Lack of semantic structure in vector quantized representations limits interpretable human control.
method Combines vector quantization with Self-Organizing Maps to learn discrete codebooks with explicit topology.
result SOM-VQ produces more learnable token sequences and provides an explicit navigable geometry in code space.
Study finds optimal learning rate schedules for sub-100M quantization-aware training across bit-widths.
problem Optimal learning rate schedules for quantization-aware training depend on bit-width.
method Factorial grid testing over bit-width, warmdown fraction, LR magnitude, model size, and seed.
result INT6 QAT requires a different schedule than higher-precision training, falsifying the primary hypothesis.
Neural network quantization is becoming an industry standard to efficiently deploy deep learning models on hardware platforms, such as CPU, GPU, TPU, and FPGAs. However, we observe that the conventional quantization approaches are vulnerable to adversarial attacks. This paper aims to raise people's awareness about the …
BatchNorm helps train quantized networks by avoiding gradient explosion.
problem Training quantized neural networks is difficult due to gradient issues.
method Investigated the impact of BatchNorm on both full-precision and quantized networks.
result BatchNorm avoids gradient explosion in quantized networks, contrary to expectations.
Optimal gradient quantization reduces communication costs in distributed deep learning.
problem High communication costs in distributed training of deep neural networks.
method Deduced optimal gradient quantization conditions for binary and multi-level quantization, developed novel schemes for dynamic quantization levels.
result Demonstrated superior performance of proposed quantization schemes on CIFAR and ImageNet datasets.
Real-time semantic segmentation for autonomous vehicles on FPGA reduces latency and power consumption.
problem Efficient real-time semantic segmentation for autonomous vehicles.
method Compressed ENet architecture, FPGA deployment, batch processing, filter reduction, quantization-aware training.
result Reduced latency to 3 ms per image with batch size of ten and 40% resource utilization.
Proposes a robust neural network quantization method.
problem Training model's dependency on specific quantization methods.
method Intrinsic robustness to various quantization processes.
result Single model capable of operating at various bit-widths and policies.
Recent machine learning methods use increasingly large deep neural networks to achieve state of the art results in various tasks. The gains in performance come at the cost of a substantial increase in computation and storage requirements. This makes real-time implementations on limited resources hardware a challenging …
New method trains quantized neural networks to global optimality.
problem Training optimal quantized neural networks is intractable due to combinatorial non-convex optimization.
method Convex optimization strategy using hidden convexity, semidefinite lifting, and Grothendieck's identity.
result Quantized NN problems can be solved to global optimality in polynomial-time.
Quantized Adam reduces communication cost in deep learning training.
problem Reducing communication cost in distributed deep learning training.
method Gradient and weight quantization with error feedback in Adam.
result Proposed methods converge to first-order stationary points.
This paper introduces a differentiable, scalable quantization method for neural networks.
problem Previous quantization methods lacked differentiability and scalability.
method The approach is differentiable and scalable, using bit-shifting and logarithmic quantization.
result The method achieves comparable accuracy to state-of-the-art approaches with less training time and lower inference cost.
Neural network quantization procedure is the necessary step for porting of neural networks to mobile devices. Quantization allows accelerating the inference, reducing memory consumption and model size. It can be performed without fine-tuning using calibration procedure (calculation of parameters necessary for quantizat…
Post-training quantization method using multiple low-precision points achieves higher precision for critical weights.
problem Discretizing pre-trained deep neural networks without re-training.
method Multipoint quantization with efficient greedy selection and adaptive point number.
result Outperforms state-of-the-art methods on ImageNet classification and PASCAL VOC object detection.
New framework reduces LLM complexity by directly finetuning in Boolean domain.
problem Reducing the complexity of large language models (LLMs) while maintaining performance.
method Proposes a novel framework using multi-kernel Boolean parameters for direct finetuning in the Boolean domain.
result Significantly reduces complexity during both finetuning and inference, outperforming recent techniques.
Currently, deep neural networks are deployed on low-power portable devices by first training a full-precision model using powerful hardware, and then deriving a corresponding low-precision model for efficient inference on such systems. However, training models directly with coarsely quantized weights is a key step towa…
The deep layers of modern neural networks extract a rather rich set of features as an input propagates through the network. This paper sets out to harvest these rich intermediate representations for quantization with minimal accuracy loss while significantly reducing the memory footprint and compute intensity of the DN…
Deep quantization of neural networks (below eight bits) offers significant promise in reducing their compute and storage cost. Albeit alluring, without special techniques for training and optimization, deep quantization results in significant accuracy loss. To further mitigate this loss, we propose a novel sinusoidal r…
Deep neural network (DNN) quantization converting floating-point (FP) data in the network to integers (INT) is an effective way to shrink the model size for memory saving and simplify the operations for compute acceleration. Recently, researches on DNN quantization develop from inference to training, laying a foundatio…
Proposes QEP to mitigate quantization error propagation in layer-wise post-training quantization.
problem Growth of quantization errors across layers degrades performance, especially in low-bit regimes.
method Quantization Error Propagation (QEP) framework that explicitly propagates and compensates for quantization errors.
result QEP-enhanced layer-wise PTQ achieves substantially higher accuracy, especially in low-bit regimes.
Neural network models are resource hungry. It is difficult to deploy such deep networks on devices with limited resources, like smart wearables, cellphones, drones, and autonomous vehicles. Low bit quantization such as binary and ternary quantization is a common approach to alleviate this resource requirements. Ternary…
Low bit-width integer weights and activations are very important for efficient inference, especially with respect to lower power consumption. We propose Monte Carlo methods to quantize the weights and activations of pre-trained neural networks without any re-training. By performing importance sampling we obtain quantiz…
AdaRound improves post-training quantization of neural networks.
problem Improving the accuracy of quantized weights in neural networks.
method Adaptive rounding mechanism that adapts to data and task loss.
result AdaRound outperforms rounding-to-nearest and achieves state-of-the-art performance.
SPEQ improves quantized neural networks by stochastic precision sharing and cosine similarity loss.
problem Improving quantized deep neural networks for edge devices.
method SPEQ combines stochastic precision sharing and cosine similarity loss for knowledge distillation.
result SPEQ outperforms existing methods in various tasks.
Recent research implies that training and inference of deep neural networks (DNN) can be computed with low precision numerical representations of the training/test data, weights and gradients without a general loss in accuracy. The benefit of such compact representations is twofold: they allow a significant reduction o…
Global-QSGD accelerates distributed training by up to 3.51%.
problem High communication overhead in distributed deep learning.
method Allreduce-compatible gradient quantization with theoretical guarantees.
result Global-QSGD accelerates distributed training by up to 3.51%.