how to visualize the detection results?

[HuangLian126]

No description provided.

[YoungXIAO13]

Hi @HuangLian126

You can run test.py with args --vis, then the visualisation is on.

[NHW2017]

Hi @HuangLian126

You can run test.py with args --vis, then the visualisation is on.

Sorry to bother you, I also wanted to see the visual results during the test, but there was an error:
File "/home/xxx/FewShotDetection-master/test.py", line 335, in
im = cv2.imread(imdb.image_path_from_index(int(data[4])))
File "/home/xxx/FewShotDetection-master/lib/datasets/pascal_voc.py", line 93, in image_path_from_index
image_path = os.path.join(self._data_path, 'JPEGImages', index + self._image_ext)
TypeError: unsupported operand type(s) for +: 'int' and 'str'
When the second error message is located, there is no problem running debug. I don’t know the reason. If you have encountered it, please let me know.

[NHW2017]

No description provided.

Did you successfully get the visual result?

[HuangLian126]

No description provided.

Did you successfully get the visual result?

No description provided.

Did you successfully get the visual result?

Yes, I successfully get the visual result. You can check your Pytorch version.

[YoungXIAO13]

Hi @NHW2017 ,

I think your problem comes from this line:
image_path = os.path.join(self._data_path, 'JPEGImages', index + self._image_ext)

Could you try modify it as:
image_path = os.path.join(self._data_path, 'JPEGImages', str(index) + self._image_ext)
to see if the issue is resolved?

[NHW2017]

Hi @NHW2017 ,

I think your problem comes from this line:
image_path = os.path.join(self._data_path, 'JPEGImages', index + self._image_ext)

Could you try modify it as:
image_path = os.path.join(self._data_path, 'JPEGImages', str(index) + self._image_ext)
to see if the issue is resolved?

No, I have tried this modification, but a new error appears and the file path cannot be found

[NHW2017]

No description provided.

Did you successfully get the visual result?

No description provided.

Did you successfully get the visual result?

Yes, I successfully get the visual result. You can check your Pytorch version.

torch==0.4.0？

[YoungXIAO13]

Then I think you might need to check the dataset is well downloaded as structured as suggested

[HuangLian126]

@NHW2017 torch==0.4, torchvision==0.21

[NHW2017]

Then I think you might need to check the dataset is well downloaded as structured as suggested

I don't quite understand, because I had no problems during training and testing, but this error occurred during --vis, so I think the error is not due to the configuration of the data set

[NHW2017]

@NHW2017 torch==0.4, torchvision==0.21

Which version of python do you use and whether you have modified other parts of the code?