- ResNeXt101_wsl: [Exploring the Limits of Weakly Supervised Pretraining](https://arxiv.org/abs/1805.00932), Dhruv Mahajan, Ross Girshick, Vignesh Ramanathan, Kaiming He, Manohar Paluri, Yixuan Li, Ashwin Bharambe, Laurens van der Maaten
- Fix_ResNeXt101_wsl: [Fixing the train-test resolution discrepancy](https://arxiv.org/abs/1906.06423), Hugo Touvron, Andrea Vedaldi, Matthijs Douze, Herve ́ Je ́gou
- EfficientNet: [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://arxiv.org/abs/1905.11946), Mingxing Tan, Quoc V. Le
- Res2Net: [Res2Net: A New Multi-scale Backbone Architecture](https://arxiv.org/abs/1904.01169), Shang-Hua Gao, Ming-Ming Cheng, Kai Zhao, Xin-Yu Zhang, Ming-Hsuan Yang, Philip Torr
- HRNet: [Deep High-Resolution Representation Learning for Visual Recognition](https://arxiv.org/abs/1908.07919), Jingdong Wang, Ke Sun, Tianheng Cheng, Borui Jiang, Chaorui Deng, Yang Zhao, Dong Liu, Yadong Mu, Mingkui Tan, Xinggang Wang, Wenyu Liu, Bin Xiao
<aname="trans1">[1]</a> The pretrained model is distilled based on the pretrained model of ResNet50_vd. Users can directly load the pretrained model through the structure of ResNet50_vd.
<aname="trans">[1]</a> means the pretrained weight is converted form [original repository](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet).
<aname="trans2">[2]</a> means the pretrained weight is converted form [original repository](https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet).
<aname="trans">[2]</a> means the pretrained weight is based on EfficientNetB0, removed Squeeze-and-Excitation module and use general convolution. This model speed is much faster.
<aname="trans3">[3]</a> means the pretrained weight is based on EfficientNetB0, removed Squeeze-and-Excitation module and use general convolution. This model speed is much faster.
@@ -579,6 +611,8 @@ Enforce failed. Expected x_dims[1] == labels_dims[1], but received x_dims[1]:100
- ResNeXt101_wsl: [Exploring the Limits of Weakly Supervised Pretraining](https://arxiv.org/abs/1805.00932), Dhruv Mahajan, Ross Girshick, Vignesh Ramanathan, Kaiming He, Manohar Paluri, Yixuan Li, Ashwin Bharambe, Laurens van der Maaten
- Fix_ResNeXt101_wsl: [Fixing the train-test resolution discrepancy](https://arxiv.org/abs/1906.06423), Hugo Touvron, Andrea Vedaldi, Matthijs Douze, Herve ́ Je ́gou
- EfficientNet: [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://arxiv.org/abs/1905.11946), Mingxing Tan, Quoc V. Le
- Res2Net: [Res2Net: A New Multi-scale Backbone Architecture](https://arxiv.org/abs/1904.01169), Shang-Hua Gao, Ming-Ming Cheng, Kai Zhao, Xin-Yu Zhang, Ming-Hsuan Yang, Philip Torr
- HRNet: [Deep High-Resolution Representation Learning for Visual Recognition](https://arxiv.org/abs/1908.07919), Jingdong Wang, Ke Sun, Tianheng Cheng, Borui Jiang, Chaorui Deng, Yang Zhao, Dong Liu, Yadong Mu, Mingkui Tan, Xinggang Wang, Wenyu Liu, Bin Xiao
## Update
...
...
@@ -593,6 +627,7 @@ Enforce failed. Expected x_dims[1] == labels_dims[1], but received x_dims[1]:100
