mmdeploy/docs/en/codebases/mmseg.md

## MMSegmentation Support

MMSegmentation is an open source object segmentation toolbox based on PyTorch. It is a part of the [OpenMMLab](https://openmmlab.com/) project.

### MMSegmentation installation tutorial

Please refer to [get_started.md](https://github.com/open-mmlab/mmsegmentation/blob/master/docs/en/get_started.md#installation) for installation.

### List of MMSegmentation models supported by MMDeploy

| Model                       | OnnxRuntime | TensorRT | NCNN | PPLNN | OpenVino |                                       Model config                                       |
|:----------------------------|:-----------:|:--------:|:----:|:-----:|:--------:|:----------------------------------------------------------------------------------------:|
| FCN                         |      Y      |    Y     |  Y   |   Y   |    Y     |      [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/fcn)      |
| PSPNet[*](#static_shape)    |      Y      |    Y     |  Y   |   Y   |    Y     |    [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/pspnet)     |
| DeepLabV3                   |      Y      |    Y     |  Y   |   Y   |    Y     |   [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/deeplabv3)   |
| DeepLabV3+                  |      Y      |    Y     |  Y   |   Y   |    Y     | [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/deeplabv3plus) |
| Fast-SCNN[*](#static_shape) |      Y      |    Y     |  N   |   Y   |    Y     |   [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/fastscnn)    |
| UNet[*](#static_shape)      |      Y      |    Y     |  Y   |   Y   |    Y     |     [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/unet)      |

### Reminder

- Only `whole` inference mode is supported for all mmseg models.

- <i id="static_shape">PSPNet, Fast-SCNN</i> only support static shape, because [nn.AdaptiveAvgPool2d](https://github.com/open-mmlab/mmsegmentation/blob/97f9670c5a4a2a3b4cfb411bcc26db16b23745f7/mmseg/models/decode_heads/psp_head.py#L38) is not supported in most of backends dynamically.

- For models only supporting static shape, you should use the deployment config file of static shape such as `configs/mmseg/segmentation_tensorrt_static-1024x2048.py`.

### FAQs

None
[Docs]: Add sphinx document (#109) * Add document structure * better zh_cn document * add build example, update requirement * add readme * add usage * fix end of file * fix codebase names, add TODO 2021-10-09 14:10:42 +08:00			`## MMSegmentation Support`
[Docs] add mmseg.md (#174) * add mmseg.md * Align with pr #171 * left alignment of 1st and last column 2021-11-08 16:08:41 +08:00
			`MMSegmentation is an open source object segmentation toolbox based on PyTorch. It is a part of the [OpenMMLab](https://openmmlab.com/) project.`

			`### MMSegmentation installation tutorial`

【Docs】Fix docs about codebase (#317) * fix codebase docs * fix edit * fix mmcls docs * fix docs 2021-12-23 17:35:59 +08:00			`Please refer to [get_started.md](https://github.com/open-mmlab/mmsegmentation/blob/master/docs/en/get_started.md#installation) for installation.`
[Docs] add mmseg.md (#174) * add mmseg.md * Align with pr #171 * left alignment of 1st and last column 2021-11-08 16:08:41 +08:00
			`### List of MMSegmentation models supported by MMDeploy`

Support UNet in mmseg (#77) * Repeatdataset in train has no CLASSES & PALETTE * update result for unet * update docstring for mmdet * remove ppl for unet in docs 2022-01-19 19:11:04 +08:00			`\| Model \| OnnxRuntime \| TensorRT \| NCNN \| PPLNN \| OpenVino \| Model config \|`
			`\|:----------------------------\|:-----------:\|:--------:\|:----:\|:-----:\|:--------:\|:----------------------------------------------------------------------------------------:\|`
			`\| FCN \| Y \| Y \| Y \| Y \| Y \| [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/fcn) \|`
			`\| PSPNet[*](#static_shape) \| Y \| Y \| Y \| Y \| Y \| [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/pspnet) \|`
			`\| DeepLabV3 \| Y \| Y \| Y \| Y \| Y \| [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/deeplabv3) \|`
			`\| DeepLabV3+ \| Y \| Y \| Y \| Y \| Y \| [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/deeplabv3plus) \|`
			`\| Fast-SCNN[*](#static_shape) \| Y \| Y \| N \| Y \| Y \| [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/fastscnn) \|`
Replace convert_syncbatchnorm in mmseg (#93) * replace convert_syncbatchnorm with revert_sync_batchnorm from mmcv * change logger 2022-01-26 11:07:28 +08:00			`\| UNet[*](#static_shape) \| Y \| Y \| Y \| Y \| Y \| [config](https://github.com/open-mmlab/mmsegmentation/tree/master/configs/unet) \|`
[Docs] add mmseg.md (#174) * add mmseg.md * Align with pr #171 * left alignment of 1st and last column 2021-11-08 16:08:41 +08:00
			`### Reminder`

[Doc]: Fix doc (#219) * add reminder for pspnet in doc * update table format * fix format * rename file * fix comment 2021-11-25 14:51:21 +08:00			- Only `whole` inference mode is supported for all mmseg models.

Add mmseg performance (#254) * update doc and config * test with ncnn * update docs * rename config * update test result with openvino * merge master * resolve comments * update dbnet 2021-12-09 20:17:00 +08:00			`- <i id="static_shape">PSPNet, Fast-SCNN</i> only support static shape, because [nn.AdaptiveAvgPool2d](https://github.com/open-mmlab/mmsegmentation/blob/97f9670c5a4a2a3b4cfb411bcc26db16b23745f7/mmseg/models/decode_heads/psp_head.py#L38) is not supported in most of backends dynamically.`

			- For models only supporting static shape, you should use the deployment config file of static shape such as `configs/mmseg/segmentation_tensorrt_static-1024x2048.py`.
[Docs] add mmseg.md (#174) * add mmseg.md * Align with pr #171 * left alignment of 1st and last column 2021-11-08 16:08:41 +08:00
			`### FAQs`

			`None`