wangxin93/fashion_compatibility_mcn

Outfit Compatibility Prediction and Diagnosis with Multi-Layered Comparison Network

fashion_compatibility_mcn's Issues

Issue in polyvore dataset trained model

I am able to train model from polyvore dataset, and after training when I run diagnosis.ipynb I get this error:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-14-59c7a44be6c0> in <module>
     10 
     11 # Step 2: show diagnosis results
---> 12 relation, out = defect_detect(x, model)
     13 relation = relation.squeeze().cpu().data
     14 show_rela_diagnosis(relation, select, cmap=plt.cm.Blues)

<ipython-input-8-e1c436475b1a> in defect_detect(img, model, normalize)
     31 
     32     if normalize:
---> 33         relation = relation / (relation.max() - relation.min())
     34     relation += 1e-3
     35     return relation, out.item()

AttributeError: 'NoneType' object has no attribute 'max'

Why is this error ?

Polyvore-T dataset

Hi Wang, can we have the dataset that you've mentioned in your paper or you could at least provide the source code to reproduce the dataset from the original Polyvore dataset?
Thanks

python main.py is not working

Fails with:
python3 main.py

Traceback (most recent call last):
  File "main.py", line 29, in <module>
    model.load_state_dict(torch.load("../mcn/model_train_relation_vse_type_cond_scales.pth", map_location="cpu"))
  File "/usr/local/lib/python3.8/site-packages/torch/serialization.py", line 571, in load
    with _open_file_like(f, 'rb') as opened_file:
  File "/usr/local/lib/python3.8/site-packages/torch/serialization.py", line 229, in _open_file_like
    return _open_file(name_or_buffer, mode)
  File "/usr/local/lib/python3.8/site-packages/torch/serialization.py", line 210, in __init__
    super(_open_file, self).__init__(open(name, mode))
FileNotFoundError: [Errno 2] No such file or directory: '../mcn/model_train_relation_vse_type_cond_scales.pth'

It looks like it tries to find /mcn/model_train_relation_vse_type_cond_scales.pth but it doesn't exist

polyvore-T unavailable

Can someone please tell me how do I get polyvore-T dataset??
the links direct me to normal polyvore but not polyvore-T

Facing 'FileNotFound' errors when diagnosing using Polyvore dataset

Hi! I cloned this repo, and as mentioned I downloaded Polyvore dataset from https://github.com/gcucurull/visual-compatibility an placed the images folder in 'fashion_compatibility_mcn/data' directory. For IDs the json files from this repo is to be used so they're in 'data' directory with names like - 'test_no_dup_with_category_3more_name.json'

So I try to execute the cells within diagnosis.ipynb but I'm facing 'FileNotFound' error at next(iterator) stage after executing prepare_dataloaders() call. Many outfit ID directories are not found.
Is there any processing of data involved with Polyvore dataset, so as to use it code from this repo?
Where am I going wrong? Please mention that and solve this issue.
Also please you make Polyvore-T dataset images available

Where is the PDF?

It is very interesting, could you share the PDF?

Making Input any number of categories

I need help in how to make the model's input to handle any number of categories in input data.

Issue in bi-LSTM for baseline

Thank you for your effort to provide well-organized code and explanation! :)
It is really helpful to understand your paper.
When I use your code for bi-LSTM, there is an issue about train.py, line 86,

• [ model._forward_and_backward(image_seqs, lengths)]

I try to find out _forward_and_backward for model but it is not existed.
How can I fix this probelm?

Thank you.

How did you split train, test, val in paper?

You mention in paper that you used "Graph segmentation is used to split train, validation and test dataset."

What exactly is that? I am trying to re-create your model training pipeline on my custom data. But the loss is not decreasing and is everywhere. I have tried everything from diagnosing image tensors, changing parameters, adjusting regularization.

I am thinking my problem might be because I randomly split the outfits in train, test and val set.

Please help me on how Graph Segmentation can be done? I am dealing in Python FYI.

TIA

Demo is not working

Demo application is not showing any output after submission.