IndexError: index out of range in SMS Text Classifier test cell

Python
1

Hi, I’m stuck on the SMS Text Classifier project. My model seems to work fine, but the testing cell throws an IndexError: index out of range.

​I suspect it’s happening when the test cell tries to access the result from my predict_message function.

​Here is my predict_message function:

def predict_message(pred_text):
prediksi = Final_model.predict(pred_text)
prediction = [prediksi]
if prediksi > 0.5:
prediction.append(“spam”)
else:
prediction.append(“ham”)

return prediction

pred_text = np.array([“big prize! £1000 cash! sale only today! call no 19282727 to stop this message”])
final = Tokenin(pred_text)
Versi_np = final.numpy().astype(np.int64)
layer_0 = Versi_np[0]
expand = np.expand_dims(layer_0, axis=0)

prediction = predict_message(expand)
print(prediction)

​And here is the full error trace:

1/1 ━━━━━━━━━━━━━━━━━━━━ 0s 56ms/step


---------------------------------------------------------------------------


IndexError                                Traceback (most recent call last)


/tmp/ipython-input-3870904172.py in <cell line: 0>()
     29   return (tensor, y_final)
     30 
---> 31 test_predictions()

3 frames

/tmp/ipython-input-3870904172.py in test_predictions()
     18 
     19   for msg, ans in zip(test_messages, test_answers):
---> 20     prediction = predict_message(msg)
     21     if prediction[1] != ans:
     22       passed = False



/tmp/ipython-input-1103236006.py in predict_message(pred_text)
      3 
      4 def predict_message(pred_text):
----> 5   prediksi = Final_model.predict(pred_text)
      6   prediction = [prediksi]
      7   if prediksi > 0.5:



/usr/local/lib/python3.12/dist-packages/keras/src/utils/traceback_utils.py in error_handler(*args, **kwargs)
    120             # To get the full stack trace, call:
    121             # `keras.config.disable_traceback_filtering()`
--> 122             raise e.with_traceback(filtered_tb) from None
    123         finally:
    124             del filtered_tb



/usr/local/lib/python3.12/dist-packages/keras/src/trainers/data_adapters/data_adapter_utils.py in <genexpr>(.0)
    103 def check_data_cardinality(data):
    104     num_samples = set(
--> 105         int(i.shape[0]) for i in tree.flatten(data) if i is not None
    106     )
    107     if len(num_samples) > 1:



IndexError: tuple index out of range

​Can anyone help me see if I’m returning the wrong format or if it’s a shape mismatch?

2

using “tree.flatten(data)” already flattened the data to a 1D data.. assuming the original array data is [[0.1,0.2,0.3], [0.4,0.5,0.6]] which is 2D, flattening it would result it to become 1D: [0.1,0.2,0.3,0.4,0.5,0.6],
Then you now tried to access it each element in the array and gettting their shape, which would result in an error.. if i can see how the original array looks like before flattening that would help better