vedantyadav/vqa

VQA - Visual Question Answering

vqa's Issues

Error when embedding

I got error when embedding

D:\Downloads\VQA-master\VQA-master>python embedding.py -address D:\Downloads\VQA-master\VQA-master\embeddings\glove.840B.300d.txt
Preparing embeddings ...
Traceback (most recent call last):
File "embedding.py", line 52, in
if name == 'main': main()
File "embedding.py", line 50, in main
create(args.address)
File "embedding.py", line 26, in create
for i, line in enumerate(f):
File "C:\Users\user\AppData\Local\Programs\Python\Python36\lib\encodings\cp1252.py", line 23, in decode
return codecs.charmap_decode(input,self.errors,decoding_table)[0]
UnicodeDecodeError: 'charmap' codec can't decode byte 0x90 in position 962: character maps to

Validation Accuracy remains constant

I trained you model for 10epochs first and then for 25 epochs
One thing I noticed is that with every epoch trainig accuracy increases but the validation accuracy remains alomost constant
and also the prediction results are unexpected.

Can you please suggest something to resolve this issue

Can you please tell me what is this file.. i am trying but i didn't get it..

features_path = 'coco_features/vgg_feats.mat'
in the prepare_data.py it seems like pre genrated..

Memory error

I am able to create embeddings but while running train.py it returns memory error. If you can share your pretrained weights it would be great help

Error while running question_answer.py

Hi,
I trained the model for few epochs and tried to test it using question_answer.py and using the coomns you provided
But it ended up with following error
ValueError: You are trying to load a weight file containing 4 layers into a model with 2 layers.

RAM of GPU used?

I tried training the model on a GPU with 10GB free space. But got memory error in the line: embedding_matrix = np.array(data)
in the file embedding.py . Can you please tell the RAM of the gpu you used to train this model.

Can you provide your built models ?

Via google drive or other platform ?