-
我使用MASKRCNN做实例分割,用的自己的数据集,这个数据集里面MASK标注是个单通道图像,每个点对应原图的实例ID,而非每个实例的轮廓多边形。我看到fluid.layers.generate_mask_labels的gt_segms参数需要是轮廓多边形,请问是否有别的方法替代generate_mask_labels?还是把轮廓线提出来,轮廓线的所有点作为gt_segms?
-
I follow the instruction and download the required data, but I get the following error after running sample_composite.sh
File
"/home/notebook/code/multiModal/Text2Img/cocoapi/PythonAPI/Text2Sce…
-
After performing the following steps:
```
git clone https://github.com/cocodataset/cocoapi.git
cd cocoapi/PythonAPI
make
```
```
I get this output error:
python setup.py build_ext --inplace
…
-
```
mona@goku:~/research/code/cocoapi/PythonAPI$ make
python setup.py build_ext --inplace
running build_ext
building 'pycocotools._mask' extension
creating build
creating build/common
creating …
-
同一批数据使用maskrcnn可以训练,使用solov2报错,2.1和rc2.0版本都是同一个错误
```
loading annotations into memory...
Done (t=0.01s)
creating index...
index created!
W0526 15:50:47.956619 30358 device_context.cc:362] Please…
-
Used in the Tiny Person Detection competition:
https://competitions.codalab.org/competitions/24551#learn_the_details-evaluation
-
I'm currently using COCOEvaluator for testing. I get AP50 and AP75 fairly easily, and it's really useful for benchmarking. However, I wonder whether it would be possible for me to change the IOU thre…
-
## ❔Question
Hi there :)
I was testing with different 'conf-thres' values and I have noticed that when setting extremely high values, the calculated
mAP@.5:.95 differs very much from the IoU=0.50:…
-
@AlexeyAB
For MS COCO test data (**test-dev**), MS did not release annotations for it. Instead, you train on the train/val datasets and submit results for the test data to the evaluation server (ht…
-
Hi, thanks for releasing the codes. I'm trying to inference the model in Youtube-VIS dataset with the trained model checkpoints you gave.
But I just got 0.286648836.
The command I used is :
`…