mmclassification/mmcls/models
Ma Zerun c708770b42
[Enhance] Support dynamic input shape for ViT-based algorithms. (#706)
* Move `resize_pos_embed` to `mmcls.models.utils`

* Refactor Vision Transformer

* Refactor DeiT

* Refactor MLP-Mixer

* Refactor Swin-Transformer

* Remove `indexing` arg

* Support dynamic inputs for t2t_vit

* Add copyright

* Fix bugs in swin transformer

* Add `pad_small_maps` option

* Update swin transformer

* Handle `attn_mask` in checkpoints of swin

* Imporve by comments
2022-03-03 13:10:12 +08:00
..
backbones [Enhance] Support dynamic input shape for ViT-based algorithms. (#706) 2022-03-03 13:10:12 +08:00
classifiers [Refactor] Remove deprecation. (#633) 2022-02-10 09:30:18 +08:00
heads
losses [Enhance] Use PyTorch official `one_hot` to implement `convert_to_one_hot`. (#696) 2022-02-28 10:20:31 +08:00
necks [Feature] Support gem pooling (#677) 2022-02-17 02:17:36 +08:00
utils [Enhance] Support dynamic input shape for ViT-based algorithms. (#706) 2022-03-03 13:10:12 +08:00
__init__.py
builder.py