PVT v2: Improved baselines with pyramid vision transformer

🤖 Plain-English Summary

Transformers have recently lead to encouraging progress in computer vision. We hope this work will facilitate advanced transformer research in computer vision.

🔑 Key Findings

• In this work, we present new baselines by improving the original Pyramid Vision Transformer (PVT v1) by adding three designs: (i) a linear complexity attention layer, (ii) an overlapping patch embedding, and (iii) a convolutional feed-forward network.
• With these modifications, PVT v2 reduces the computational complexity of PVT v1 to linearity and provides significant improvements on fundamental vision tasks such as classification, detection, and segmentation.
• In particular, PVT v2 achieves comparable or better performance than recent work such as the Swin transformer.

🔗 Related Resources