search

yunjey / pytorch-tutorial

PyTorch Tutorial for Deep Learning Researchers
MIT License
29.79k stars 8.03k forks source link

[image caption] Test error when encoder(image) #152

Open apeterswu opened 5 years ago

apeterswu commented 5 years ago

Hi,

I met the problem when I tried to decode a caption for images from val2014 dataset, for example, the picture is COCO_val2014_000000007888.jpg The error happened when processing feature = encoder(image)

RuntimeError: Given groups=1, weight[64, 3, 7, 7], so expected input[1, 1, 224, 224] to have 3 channels, but got 1 channels instead

Is there anything wrong for the encoder processing? Or any transformation need to be done for the picture? I use the sample.py script. Thanks.

OswaldoBornemann commented 5 years ago

same problem, have you solved it ?

mdhasanai commented 5 years ago

I got the same problem.. and i solved it.. When you load image, make sure that the image has three color channel (RGB) because it might be gray scale. So you should convert it. An example is below:

Here is my code,

from PIL import Image

img = Image.open(image_path).convert('RGB')