Question

我是一个初学者，正在学习编码图像分类器。我的目标是创建一个预测函数。

有什么建议可以解决吗？

在这个项目中，我想使用预测功能来识别不同的花卉种类。所以我以后可以检查他们的标签。

试图修复：不幸的是，该错误仍然持续存在。我已经尝试过这些代码：

img = process_image(Image.open(image))
img = torch.from_numpy(img).type(torch.FloatTensor)

这是我现在需要修复的错误。

AttributeError：'JpegImageFile'对象没有属性'read'

代码：

# Imports here
import pandas as pd
import numpy as np

import torch
from torch import nn
from torchvision import datasets, transforms, models
import torchvision.models as models
import torch.nn.functional as F
import torchvision.transforms.functional as F
from torch import optim
import json

from collections import OrderedDict
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
%matplotlib inline
from PIL import Image

def predict(image, model, topk=5):

#Predict the class (or classes) of an image using a trained deep #learning model.
#Here, image is the path to an image file, but input to process_image #should be Image.open(image).                                                         

    img = process_image(Image.open(image))
    img = torch.from_numpy(img).type(torch.FloatTensor) 

    output = model.forward(img)
    probs, labels = torch.topk(output, topk)        
    probs = probs.exp()

    # Reverse the dict
    idx_to_class = {val: key for key, val in
model.class_to_idx.items()}
    # Get the correct indices
    top_classes = [idx_to_class[each] for each in classes]

    return labels, probs

predict(image,model)
print(probs)
print(classes)

错误：

AttributeError                            Traceback (most recent call last)
<ipython-input-32-b49fdcab5791> in <module>()
----> 1 probs, classes = predict(image, model)
      2 print(probs)
      3 print(classes)

<ipython-input-31-6f996290ea63> in predict(image, model, topk)
      5       Image.open(image)
      6     '''
----> 7     img = process_image(Image.open(image))
      8     img = torch.from_numpy(img).type(torch.FloatTensor)
      9 

/opt/conda/lib/python3.6/site-packages/PIL/Image.py in open(fp, mode)
   2587         exclusive_fp = True
   2588 
-> 2589     prefix = fp.read(16)
   2590 
   2591     preinit()

AttributeError: 'JpegImageFile' object has no attribute 'read'

我要做的就是得到这些相似的结果。谢谢！

tensor([[ 0.5607,  0.3446,  0.0552,  0.0227,  0.0054]], device='cuda:0')   
tensor([[  8,   1,  31,  24,   7]], device='cuda:0')

Answer 1

尝试使用函数YOLOv3 darknet53.conv.74 obj.data obj.names Tank.zip yolov3.weights yolov3_custom.cfg yolov3_custom1.cfg.txt将JpegImageFile转换为numpy数组，然后就可以使用图像了。

Answer 2

什么是image？它来自哪里？

似乎已经Image.open()编辑过，您正尝试重新读取它，就像它是文件还是路径一样。

Answer 3

尝试打开已调用Image.open（）的图像时遇到相同的问题。以下将为您工作：

img = process_image(image)
img = torch.from_numpy(img).type(torch.FloatTensor)

如果图像的类型和/或图像，您将看到它已经是张量：

print(image
print(type(image))

这是一个回答相同问题的相关链接：

https://discuss.pytorch.org/t/typeerror-img-should-be-pil-image-got-class-str/49644

Answer 4

打开后将图像转换为RGB为我解决了这个问题

   PIL.Image.open(image_path).convert('RGB')

Answer 5

我遇到了类似的问题，通过如下打开图片解决了：

import io
from PIL import Image as pil_image

fname = 'my_image.jpg'

with open(fname, 'rb') as f:
    img = pil_image.open(io.BytesIO(f.read()))
    pass

而不是：

from PIL import Image as pil_image

fname = 'my_image.jpg'

with pil_image.open(fname, 'r') as img:
    pass

如何解决“ AttributeError：'JpegImageFile'对象没有属性'read'？

5 个答案: