[FIXED] TypeError: unhashable type: 'numpy.ndarray' Python3.9 image classification using tensorflow and keras

Issue

I try this sample code for image classification

def show_classify_button(file_path):
    classify_btn = Button(top, text="Classify Image", command=lambda: classify(file_path), padx=10, pady=5)
    classify_btn.configure(background="#364156", foreground="white", font=('arial',10,'bold'))
    classify_btn.place(relx=0.79,rely=0.46)

def classify(file_path):
    image = Image.open(file_path)
    image = image.resize((32,32))
    image = numpy.expand_dims(image, axis=0)
    image = numpy.array(image)
    pred = model.predict([image])[0]
    sign = classes[pred]
    print(sign)
    label.configure(foreground='#011638')

the terminal pop this

Traceback (most recent call last):
  line 39, in <lambda>
    classify_btn = Button(top, text="Classify Image", command=lambda: classify(file_path), padx=10, pady=5)
  line 49, in classify
    sign = classes[pred]
TypeError: unhashable type: 'numpy.ndarray'

I try to check the data from the pred with output

[30990.06  46435.57  17636.973 16334.658 15860.342 16765.371 26879.748
 14579.97  41989.523 34359.215]

im not sure why because the data is from set of an array

im new with this and im using python3.9 can someone help me

Solution

You're trying to access classes variable on line 49

sign = classes[pred]

classes is of type numpy.ndarray. So you're trying to access an array at index pred but because pred is not an number it's raising a unhashable type: 'numpy.ndarray' error.

You're treating classes like a dictionary by accessing it's values with a key and not with an index.

Answered By - Hack_Hut