Boost Vision Transformer with GPU-Friendly Sparsity and Quantization
Yu, Chong, Chen, Tao, Gan, Zhongxue, Fan, Jiayuan
–arXiv.org Artificial Intelligence
The transformer extends its success from the language to the vision domain. Because of the stacked self-attention and cross-attention blocks, the acceleration deployment of vision transformer on GPU hardware is challenging and also rarely studied. This paper thoroughly designs a compression scheme to maximally utilize the GPU-friendly 2:4 fine-grained structured sparsity and quantization. Specially, an original large model with dense weight parameters is first pruned into a sparse one by 2:4 structured pruning, which considers the GPU's acceleration of 2:4 structured sparse pattern with FP16 data type, then the floating-point sparse model is further quantized into a fixed-point one by sparse-distillation-aware quantization aware training, which considers GPU can provide an extra speedup of 2:4 sparse calculation with integer tensors. A mixed-strategy knowledge distillation is used during the pruning and quantization process. The proposed compression scheme is flexible to support supervised and unsupervised learning styles. Experiment results show GPUSQ-ViT scheme achieves state-of-the-art compression by reducing vision transformer models 6.4-12.7 times on model size and 30.3-62 times on FLOPs with negligible accuracy degradation on ImageNet classification, COCO detection and ADE20K segmentation benchmarking tasks. Moreover, GPUSQ-ViT can boost actual deployment performance by 1.39-1.79 times and 3.22-3.43 times of latency and throughput on A100 GPU, and 1.57-1.69 times and 2.11-2.51 times improvement of latency and throughput on AGX Orin.
arXiv.org Artificial Intelligence
May-18-2023
- Genre:
- Research Report (0.70)
- Industry:
- Information Technology (0.48)
- Technology:
- Information Technology
- Artificial Intelligence
- Machine Learning > Neural Networks
- Deep Learning (0.68)
- Vision (1.00)
- Machine Learning > Neural Networks
- Graphics (1.00)
- Hardware (1.00)
- Artificial Intelligence
- Information Technology