MEAL V2: Boosting Vanilla ResNet-50 to 80%+ Top-1 Accuracy on ImageNet without Tricks

Abstract

We introduce a simple yet effective distillation framework that is able to boost the vanilla ResNet-50 to 80%+ Top-1 accuracy on ImageNet without tricks. We construct such a framework through analyzing the problems in the existing classification system and simplify the base method ensemble knowledge distillation via discriminators by: (1) adopting the similarity loss and discriminator only on the final outputs and (2) using the average of softmax probabilities from all teacher ensembles as the stronger supervision. Intriguingly, three novel perspectives are presented for distillation: (1) weight decay can be weakened or even completely removed since the soft label also has a regularization effect; (2) using a good initialization for students is critical; and (3) one-hot/hard label is not necessary in the distillation process if the weights are well initialized. We show that such a straight-forward framework can achieve state-of-the-art results without involving any commonly-used techniques, such as architecture modification; outside training data beyond ImageNet; autoaug/randaug; cosine learning rate; mixup/cutmix training; label smoothing; etc. Our method obtains 80.67% top-1 accuracy on ImageNet using a single crop-size of 224x224 with vanilla ResNet-50, outperforming the previous state-of-the-arts by a significant margin under the same network structure. Our result can be regarded as a strong baseline using knowledge distillation, and to our best knowledge, this is also the first method that is able to boost vanilla ResNet-50 to surpass 80% on ImageNet without architecture modification or additional training data. On smaller ResNet-18, our distillation framework consistently improves from 69.76% to 73.19%, which shows tremendous practical values in real-world applications. Our code and models are available at: https://github.com/szq0214/MEAL-V2.

12 pages. Code and trained models are available at: https://github.com/szq0214/MEAL-V2

References in corpus (9)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• Distilling the Knowledge in a Neural Network
• Neural Architecture Search with Reinforcement Learning
• Improved Regularization of Convolutional Neural Networks with Cutout
• FitNets: Hints for Thin Deep Nets
• Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
• Trained Ternary Quantization
• Pruning Filters for Efficient ConvNets
• Is Label Smoothing Truly Incompatible with Knowledge Distillation: An Empirical Study

Cited by in corpus (11)

• SEED: Self-supervised Distillation For Visual Representation
• Towards Understanding and Boosting Adversarial Transferability from a Distribution Perspective
• Is Label Smoothing Truly Incompatible with Knowledge Distillation: An Empirical Study
• Joint Multi-Dimension Pruning via Numerical Gradient Update
• NPAS: A Compiler-aware Framework of Unified Network Pruning and Architecture Search for Beyond Real-Time Mobile Acceleration
• Student Network Learning via Evolutionary Knowledge Distillation
• DS-Net++: Dynamic Weight Slicing for Efficient Inference in CNNs and Transformers
• S2-BNN: Bridging the Gap Between Self-Supervised Real and 1-bit Neural Networks via Guided Distribution Calibration
• A Fast Knowledge Distillation Framework for Visual Recognition
• Arch-Net: Model Distillation for Architecture Agnostic Model Deployment
• Dynamic Slimmable Network