PascalWB
commented
6 years ago For anyone facing issues with the tutorial:
There are a couple of code lines missing. You have to run the session before plotting the images.
Insert:
Run and get result image
preds = sess.run(pred)
data = sess.run(img_np)just before plotting.
Thank you for all the input. I have, however, problems to get the tutorial (https://modeldepot.io/hellochick/pspnet) running on my machine. Unfortunately, I'm very new to python.
Minimal example: import tensorflow as tf import numpy as np from scipy import misc import matplotlib.pyplot as plt
from PIL import Image
from model import PSPNet101, PSPNet50 from tools import *
image_path = './input/test1.png'
img_np, filename = load_img(image_path) plt.imshow(img_np)
This leads to an error message:
_img_np, filename = load_img(image_path) plt.imshow(img_np) successful load img: ./input/test1.png Traceback (most recent call last):
File "", line 2, in
plt.imshow(img_np)
File "C:\Anaconda3\lib\site-packages\matplotlib\pyplot.py", line 3080, in imshow **kwargs)
File "C:\Anaconda3\lib\site-packages\matplotlib__init__.py", line 1710, in inner return func(ax, *args, **kwargs)
File "C:\Anaconda3\lib\site-packages\matplotlib\axes_axes.py", line 5194, in imshow im.set_data(X)
File "C:\Anaconda3\lib\site-packages\matplotlib\image.py", line 600, in set_data raise TypeError("Image data cannot be converted to float")
TypeError: Image data cannot be converted to float_
The problem seems to be that img_np is shapeless: _imgnp Out[3]: <tf.Tensor 'DecodePng:0' shape=(?, ?, 3) dtype=uint8>
This issue is well discussed (https://github.com/tensorflow/tensorflow/issues/9356), but I could not figure out the solution - does anybody has a hint?
I'd like to read in images of different sizes.
All the best, Pascal