can you guide me to change the code so that the code is also applicable for trying in pants. Please help🙏

Hi @lastdefiance20 there seems to be lot of noise in the output and full sleeves and even sleeveless t-shirts often gets their sleeves short? what could be the issue? Is it that dependencies may be the problems. Why is it giving half sleeves all the time?
Is there any parameter which i can tweak to get better results?
Though this is the issue while running locally.

Hi,
@lastdefiance20 I am not able to generate data.json file. Instead of getting output saved to data.json, i get output saved to output.pkl and getting the distorted image.

Can you please share your environments details for CUDA, cuDNN versions because openpose does not seem to install on colab
and giving errors like these CMU-Perceptual-Computing-Lab/openpose#1322

getting the following error please help me out

C:\Users\mhash\Downloads\TryYours-Virtual-Try-On-main2\TryYours-Virtual-Try-On-main>python main.py
Get mask of cloth

Get openpose coordinate using posenet

Traceback (most recent call last):
File "C:\Users\mhash\Downloads\TryYours-Virtual-Try-On-main2\TryYours-Virtual-Try-On-main\posenet.py", line 10, in
net = net.cuda()
^^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\nn\modules\module.py", line 915, in cuda
return self._apply(lambda t: t.cuda(device))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\nn\modules\module.py", line 779, in _apply
module._apply(fn)
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\nn\modules\module.py", line 779, in _apply
module._apply(fn)
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\nn\modules\module.py", line 779, in _apply
module._apply(fn)
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\nn\modules\module.py", line 804, in _apply
param_applied = fn(param)
^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\nn\modules\module.py", line 915, in
return self.apply(lambda t: t.cuda(device))
^^^^^^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\cuda_init.py", line 284, in _lazy_init
raise AssertionError("Torch not compiled with CUDA enabled")
AssertionError: Torch not compiled with CUDA enabled
Generate semantic segmentation using Graphonomy-Master library

Constructing DeepLabv3+ model...
Number of classes: 20
Output stride: 16
Number of Input Channels: 3
Traceback (most recent call last):
File "C:\Users\mhash\Downloads\TryYours-Virtual-Try-On-main2\TryYours-Virtual-Try-On-main\Graphonomy-master\exp\inference\inference.py", line 191, in
x = torch.load(opts.loadmodel)
^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\serialization.py", line 997, in load
with _open_file_like(f, 'rb') as opened_file:
^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\serialization.py", line 444, in _open_file_like
return open_file(name_or_buffer, mode)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\mhash\anaconda3\Lib\site-packages\torch\serialization.py", line 425, in init
super().init(open(name, mode))
^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: './inference.pth'
[ WARN:[email protected]] global loadsave.cpp:248 cv::findDecoder imread('./resized_segmentation_img.png'): can't open/read file: check file path/integrity
Traceback (most recent call last):
File "C:\Users\mhash\Downloads\TryYours-Virtual-Try-On-main2\TryYours-Virtual-Try-On-main\main.py", line 44, in
mask_img=cv2.resize(mask_img,(768,1024))
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
cv2.error: OpenCV(4.9.0) D:\a\opencv-python\opencv-python\opencv\modules\imgproc\src\resize.cpp:4152: error: (-215:Assertion failed) !ssize.empty() in function 'cv::resize'

When I upload custom cloth, it's not working. Only showing the predefined cloth.

Does this works with the custom data as well?

On which tool have you used it, did you do it on Google Colab or in Ubuntu.
Give all the steps to run it

Can you please tell what did the training data images consist of like what all images you fed into training i want to train it for different dataset of images for my project .Please Help 🙏

Hi,
This is amazing!!.. Its working superb with different images i tried. Now, Can i get any guidance on how can i make a demo UI such as gradio application or streamlit application with this on VS code. Or what essential changes will i have to make in the code files for it?

Hi,in get_densepose.py, it needs a data.json file. I can't find it or how can I generate it?

This is my input.

I got output like this.

Can anyone tell what should I change in my code ?

https://colab.research.google.com/drive/1fP5Wia4ukTp6WC5FlSa7InW7cOLXCePy?usp=sharing#scrollTo=gJUHNGLorsAR

Hi bro! In the part of Generate semantic segmentation using Graphonomy-Master library of "main.py", there is a model path named inference.pth, but i don't know how to get it or generate it, would you please tell me some infomations about it? thanks

I upload a t-shirt image, the preview is OK:

but the output final image are always the Yale T-shirt

I tried using the repo with changes in detectron2 and main.py suggested in issue@
And still getting the following error:

Get mask of cloth

Get openpose coordinate using posenet

Segmentation fault
Generate semantic segmentation using Graphonomy-Master library

Constructing DeepLabv3+ model...
Number of classes: 20
Output stride: 16
Number of Input Channels: 3
load model: ./inference.pth
Segmentation fault

Generate Densepose image using detectron2 library

[05/23 07:54:23 apply_net]: Loading config from detectron2/projects/DensePose/configs/densepose_rcnn_R_50_FPN_s1x.yaml
[05/23 07:54:23 apply_net]: Loading model from https://dl.fbaipublicfiles.com/densepose/densepose_rcnn_R_50_FPN_s1x/165712039/model_final_162be9.pkl
[05/23 07:54:25 apply_net]: Loading data from origin.jpg
Segmentation fault
Traceback (most recent call last):
File "/home/deepk/Desktop/backend/virtual_tryon/TryYours-Virtual-Try-On/get_densepose.py", line 27, in
with open('./data.json', 'r') as f:
FileNotFoundError: [Errno 2] No such file or directory: './data.json'

Run HR-VITON to generate final image

Start to test! - HR-VITON
Network [SPADEGenerator] was created. Total number of parameters: 100.5 million. To see the architecture, do print(network).
Segmentation fault