ins_seg分支YOLOv5、v8和PPYOLOE seg目前支持训练了吗 about paddleyolo

[nemonameless]

PPYOLOE seg可训，其余暂不可训。暂未支持Fastdeploy。

[panp4n]

PPYOLOE seg可训，其余暂不可训。暂未支持Fastdeploy。

我尝试了PPYOLOE-seg（640640）训练自己的数据集，检测效果良好，但分割掩膜有明显的锯齿，尤其是小目标，且评估segm, mAP异常。
同样的数据集使用MaskRCNN（8001280）检测分割效果要好的多，且segm mAP正常。
以下为PPYOLOE-seg验证Log:

DONE (t=0.01s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type bbox
DONE (t=0.23s).
Accumulating evaluation results...
DONE (t=0.03s).
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.716
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.909
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.816
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.224
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.664
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.762
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.103
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.548
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.787
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.350
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.748
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.817
[12/15 17:56:40] ppdet.metrics.metrics INFO: The mask result is saved to mask.json.
loading annotations into memory...
Done (t=0.00s)
creating index...
index created!
[12/15 17:56:40] ppdet.metrics.coco_utils INFO: Start evaluate...
Loading and preparing results...
DONE (t=0.03s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type segm
DONE (t=0.28s).
Accumulating evaluation results...
DONE (t=0.02s).
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.014
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.025
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.017
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.001
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.006
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.017
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.013
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.053
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.054
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.005
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.056
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.049
[12/15 17:56:40] ppdet.engine INFO: Total sample number: 5, average FPS: 0.500295790076311
[12/15 17:56:40] ppdet.engine INFO: Best test bbox ap is 0.718.
看预测的结果也挺正常，得分都挺高。

另外我发现训练PaddleYOLO中这几个实例分割，显存占用异常大，PPYOLOE-seg-s 640输入尺寸，2 batchsize就要占12G显存。其他YOLOv5、6、8 seg 12G显存只能跑1个batchsize。我在使用YOLOv5-seg\YOLOv8-seg官方代码训练时，可以使用更大的batchsize和更大的输入尺寸（1280），也不会爆显存，且检测和分割效果要好很多。

[nemonameless]

显存问题确实存在，后续有空会继续排查，谢谢