Layer-adaptive Structured Pruning Guided by Latency

Structured pruning can simplify network architecture and improve inference speed. Combined with the underlying hardware and inference engine in which the final model is deployed, better results can be obtained by using latency collaborative loss function to guide network pruning together. Existing pruning methods that optimize latency have demonstrated leading performance, however, they often overlook the hardware features and connection in the network. To address this problem, we propose a global importance score SP-LAMP(Structured Pruning Layer-Adaptive Magnitude-based Pruning) by deriving a global importance score LAMP from unstructured pruning to structured pruning. In SP-LAMP, each layer includes a filter with an SP-LAMP score of 1, and the remaining filters are grouped. We utilize a group knapsack solver to maximize the SP-LAMP score under latency constraints. In addition, we improve the strategy of collect the latency to make it more accurate. In particular, for ResNet50/ResNet18 on ImageNet and CIFAR10, SP-LAMP is 1.28x/8.45x faster with +1.7%/-1.57% top-1 accuracy changed, respectively. Experimental results in ResNet56 on CIFAR10 demonstrate that our algorithm achieves lower latency compared to alternative approaches while ensuring accuracy and FLOPs.

Augmentation-Aware Self-Supervision for Data-Efficient GAN Training

Training generative adversarial networks (GANs) with limited data is valuable but challenging because discriminators are prone to over-fitting in such situations. Recently proposed differentiable data augmentation techniques for discriminators demonstrate improved data efficiency of training GANs. However, the naive data augmentation introduces undesired invariance to augmentation into the discriminator. The invariance may degrade the representation learning ability of the discriminator, thereby affecting the generative modeling performance of the generator. To mitigate the invariance while inheriting the benefits of data augmentation, we propose a novel augmentation-aware self-supervised discriminator that predicts the parameter of augmentation given the augmented and original data. Moreover, the prediction task is required to distinguishable between real data and generated data since they are different during training. We further encourage the generator to learn from the proposed discriminator by generating augmentation-predictable real data. We compare the proposed method with state-of-the-arts across the class-conditional BigGAN and unconditional StyleGAN2 architectures on CIFAR-10/100 and several low-shot datasets, respectively. Experimental results show a significantly improved generation performance of our method over competing methods for training data-efficient GANs.

Fine-tuning Pruned Networks with Linear Over-parameterization

Structured pruning compresses neural networks by reducing channels (filters) for fast inference and low footprint at run-time. To restore accuracy after pruning, fine-tuning is usually applied to pruned networks. However, too few remaining parameters in pruned networks inevitably bring a great challenge to fine-tuning to restore accuracy. To address this challenge, we propose a novel method that first linearly over-parameterizes the compact layers in pruned networks to enlarge the number of fine-tuning parameters and then re-parameterizes them to the original layers after fine-tuning. Specifically, we equivalently expand the convolution/linear layer with several consecutive convolution/linear layers that do not alter the current output feature maps. Furthermore, we utilize similarity-preserving knowledge distillation that encourages the over-parameterized block to learn the immediate data-to-data similarities of the corresponding dense layer to maintain its feature learning ability. The proposed method is comprehensively evaluated on CIFAR-10 and ImageNet which significantly outperforms the vanilla fine-tuning strategy, especially for large pruning ratio.