wait for mmdet 3.0.0rc4

configs/maskformer/README.md

@@ -38,29 +38,9 @@ Modern approaches typically formulate semantic segmentation as a per-pixel class
 - MaskFormer model needs to install [MMDetection](https://github.com/open-mmlab/mmdetection) first.
 
 ```shell
-pip install "mmdet>=3.0.0rc3"
+pip install "mmdet>=3.0.0rc4"
 ```
 
-If related MMDetection version unfounded, you can modify [related code](https://github.com/open-mmlab/mmdetection/blob/dev-3.x/mmdet/models/dense_heads/maskformer_head.py#L106) like [MMDetection PR](https://github.com/open-mmlab/mmdetection/pull/9176) on your own to fix its bug:
-
-From
-
-```python
-if pixel_decoder_type == 'PixelDecoder' and (
-        self.decoder_embed_dims != in_channels[-1]
-        or enforce_decoder_input_project):
-```
-
-to
-
-```python
-if type(self.pixel_decoder) == PixelDecoder and (
-        self.decoder_embed_dims != in_channels[-1]
-        or enforce_decoder_input_project):
-```
-
-Which would treat `mmdet.PixelDecoder` the same with `PixelDecoder` in this MMDetection `if` conditional statement.
-
 ## Results and models
 
 ### ADE20K