相关论文: Ultra Memory-Efficient On-FPGA Training of Transfo…

Transformers have excelled in many tasks including vision. However, efficient deployment of transformer models in low-latency or high-throughput applications is hindered by the computation in the attention mechanism which involves expensive…

The growth rate of the GPU memory capacity has not been able to keep up with that of the size of large language models (LLMs), hindering the model training process. In particular, activations -- the intermediate tensors produced during…

The attention mechanism is a key computing kernel of Transformers, calculating pairwise correlations across the entire input sequence. The computing complexity and frequent memory access in computing self-attention put a huge burden on the…

As training billion-scale transformers becomes increasingly common, employing multiple distributed GPUs along with parallel training methods has become a standard practice. However, existing transformer designs suffer from significant…

The growing demands of distributed learning on resource constrained edge devices underscore the importance of efficient on device model compression. Tensor Train Decomposition (TTD) offers high compression ratios with minimal accuracy loss,…

Neural network accelerators with low latency and low energy consumption are desirable for edge computing. To create such accelerators, we propose a design flow for accelerating the extremely low bit-width neural network (ELB-NN) in embedded…

In Federated Learning (FL), devices that participate in the training usually have heterogeneous resources, i.e., energy availability. In current deployments of FL, devices that do not fulfill certain hardware requirements are often dropped…

Recently, large models, such as Vision Transformer and BERT, have garnered significant attention due to their exceptional performance. However, their extensive computational requirements lead to considerable power and hardware resource…

Diffusion models have recently gained state of the art performance on many image generation tasks. However, most models require significant computational resources to achieve this. This becomes apparent in the application of medical image…

Transformer is a deep learning language model widely used for natural language processing (NLP) services in datacenters. Among transformer models, Generative Pre-trained Transformer (GPT) has achieved remarkable performance in text…

Convolutional neural networks (CNNs) are revolutionizing machine learning, but they present significant computational challenges. Recently, many FPGA-based accelerators have been proposed to improve the performance and efficiency of CNNs.…

Transformers lack an explicit architectural mechanism for storing and organizing knowledge acquired during training. We introduce learnable sparse memory banks: a set of latent tokens, randomly initialized and trained end-to-end, that…

Using fewer bits to represent model parameters and related tensors during pre-training has become a required technique for improving GPU efficiency without sacrificing accuracy. Microscaling (MX) formats introduced in NVIDIA Blackwell…

In video generation models, particularly world models, training large-scale video diffusion Transformers (such as DiT and MMDiT) poses significant computational challenges due to the extreme variance in sequence lengths within mixed-mode…

Efficient execution of deep learning workloads on dataflow architectures is crucial for overcoming memory bottlenecks and maximizing performance. While streaming intermediate results between computation kernels can significantly improve…

Tucker decomposition is one of the SOTA CNN model compression techniques. However, unlike the FLOPs reduction, we observe very limited inference time reduction with Tucker-compressed models using existing GPU software such as cuDNN. To this…

The energy efficiency of neuromorphic hardware is greatly affected by the energy of storing, accessing, and updating synaptic parameters. Various methods of memory organisation targeting energy-efficient digital accelerators have been…

This paper presents the design of a hardware accelerator for Transformers, optimized for on-device time-series forecasting in AIoT systems. It integrates integer-only quantization and Quantization-Aware Training with optimized hardware…

Computationally intensive deep neural networks (DNNs) are well-suited to run on GPUs, but newly developed algorithms usually require the heavily optimized DNN routines to work efficiently, and this problem could be even more difficult for…

Deep-learning-based compressor has received interests recently due to much improved compression ratio. However, modern approaches suffer from long execution time. To ease this problem, this paper targets on cutting down the execution time…