Test with my own data

[woo1]

How can I test with my own data?
I want to test with my image and make videos like yours.
("https://github.com/tengteng95/Pose-Transfer/raw/master/imgs/dance3.gif")
I read README file but the description is only about the datasets (DeepFashion, Market-1501).
Can you give me more details?

[ElToro13]

Did you fiqure it out? I am stuck at the same point. No way to generate annotation files.

[PalakHarwani]

Hi,
I have successfully tested with my own data. You can generate annotation files by changing the paths in tool/compute_coordinates.py .

[rtx98]

@PalakHarwani can you please tell me your keras and tensorflow version? I am having trouble running the script.

[CodeAunt]

Hi, I have successfully tested with my own data. You can generate annotation files by changing the paths in tool/compute_coordinates.py .

Hi, I made it to generate the annotation csv file, but all the results are -1.
It looks like this:
fashionWOMENTees_Tanksid0000221902_3back.jpg: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1] fashionWOMENTees_Tanksid0000217201_1front.jpg: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1] fashionWOMENTees_Tanksid0000250305_1front.jpg: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1] fashionWOMENTees_Tanksid0000218703_4full.jpg: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]: [-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1]
I'm using images from deepfashion high_res, and python3.7.
Is the file 'pose_estimator.h5' valid? Or are there any other requirements to the image files? Thanks