Error in C2W2_Assignment for Feature Engineering

Course Q&A Machine Learning Engineering for Production(MLOps)
1

I get always that Error:

TypeError: Failed to convert object of type <class ‘tensorflow.python.framework.sparse_tensor.SparseTensor’> to Tensor. Contents: SparseTensor(indices=Tensor(“F_traffic_volume/Placeholder:0”, shape=(None, 2), dtype=int64), values=Tensor(“Cast:0”, shape=(None,), dtype=float32), dense_shape=Tensor(“F_traffic_volume/Placeholder_2:0”, shape=(2,), dtype=int64)). Consider casting elements to a supported type.

That is what I code:

# Keep as is. No tft function needed.
for key in _CATEGORICAL_FEATURE_KEYS:
    outputs[_transformed_name(key)] = None
    
# Use `tf.cast` to cast the label key to float32 and fill in the missing values.
traffic_volume = tf.cast(inputs[_VOLUME_KEY], tf.float32)
# Create a feature that shows if the traffic volume is greater than the mean and cast to an int
outputs[_transformed_name(_VOLUME_KEY)] = tf.cast(  
    
    # Use `tf.greater` to check if the traffic volume in a row is greater than the mean of the entire traffic volumn column
    tf.greater(traffic_volume, tft.mean(traffic_volume)), tf.int64)
2

Hello, I think you need to recheck the “_fill_in_missing” function (read its doc string) and how it’s related to the error you are getting.

2 Likes
3

Hi! You’re on the right track. To resolve this, let’s go back a bit to the blank exercise. The last three lines of the boilerplate should look like:

    outputs[_transformed_name(_VOLUME_KEY)] = tf.cast(  
        
        # Use `tf.greater` to check if the traffic volume in a row is greater than the mean of the entire traffic volumn column
        tf.greater(None, None(tf.cast(inputs[_VOLUME_KEY], tf.float32))),
        
        tf.int64)

You should only modify the second line and only the None placeholders. inputs[_VOLUME_KEY] should be cast into a float. Then the result of tf.greater will be cast into tf.int64. Hope this helps!

4

In addition to what Chris said, I don’t think there is a function call tft.mean , it should be another tf method

5

I am stuck at exactly the same place. Can you point me to something that documents how to get the mean? I can’t see anywhere that a specific statistics item (mean, min, max, …) is accessed. I see tft.scale_to_z_score which uses the mean. I can also read the mean for traffic volume from the output in section 2.2. Any pointers would be appreciated.
thanks

6

I may have solved the issue by doing exactly what tim86 said NOT to do - use tft.mean.
All seems to work now.
???