why am I getting warning/error when working with tensorflow (use functional API and not implemented error)

I am trying to follow this tutorial but with my data: https://www.tensorflow.org/tutorials/structured_data/feature_columns

All of my data is numerical values.

when I ran this part of code:

model.compile(optimizer='adam',                loss='binary_crossentropy',                metrics=['accuracy'])      history = model.fit(train_ds, validation_data=test_ds, epochs=100, use_multiprocessing=True)  

I am getting this type of warning for all of the parameters:

WARNING:tensorflow:Layers in a Sequential model should only have a single input tensor, but we receive a <class 'dict'> input: {'age': <tf.Tensor 'ExpandDims_8:0' shape=(None, 1) dtype=int64>,  

I am getting this warning twice for each variable!

and then I am getting this error:

UnimplementedError:  Cast string to float is not supported   [[node sequential_7/dense_features_7/calprotectin/Cast (defined at <ipython-input-103-5689ba5df442>:5) ]] [Op:__inference_train_function_4860]  

What is the problem and how can I fix it?

Edit1

I tried to mimic my code and error using sample data and I came up with this code.

The code doesn't generate an error but generates a warning. so the problem is with the data that I am reading. What can go wrong with the input data that generate such an error?

( it is a jupyter code, how can I post it here?) :

%reset  import numpy as np  import pandas as pd  import tensorflow as tf  from tensorflow import keras  from tensorflow import feature_column  from sklearn.model_selection import train_test_split    RANDOM_SEED = 42    data=pd.DataFrame()  data['sex']=[1,2,2,1,2,2,1,1,2,1]  data['age']=[10,11,13,45,67,34,23,62,82,78]  data['bmi']=[22.5,28.8,19,23.3,26,18.4,27.5,29,30.3,25.9]  data['smoker']=[1,2,2,3,3,2,2,1,1,1]  data['lab1']=[144,124,126,146,130,124,171,147,131,138]  data['lab2']=[71,82,75,65,56,89,55,74,78,69]  data['result']=[1,2,2,4,3,2,1,3,2,4]    feature_columns = []  for header in ['sex','age', 'bmi','smoker', 'lab1', 'lab2']:    feature_columns.append(tf.feature_column.numeric_column(header))    def create_dataset(dataframe, batch_size=32):      dataframe = dataframe.copy()      labels = dataframe.pop('result')      return tf.data.Dataset.from_tensor_slices((dict(dataframe), labels)) \        .shuffle(buffer_size=len(dataframe)) \        .batch(batch_size)    train, test = train_test_split(data, test_size=0.2, random_state=RANDOM_SEED)  train_ds = create_dataset(train)  test_ds = create_dataset(test)    model = tf.keras.models.Sequential([    tf.keras.layers.DenseFeatures(feature_columns=feature_columns),    tf.keras.layers.Dense(128, activation='relu'),    tf.keras.layers.Dense(128, activation='relu'),    tf.keras.layers.Dropout(.1),    tf.keras.layers.Dense(1)  ])    model.compile(optimizer='adam',            loss='binary_crossentropy',            metrics=['accuracy'])    history = model.fit(train_ds, validation_data=test_ds, epochs=100, use_multiprocessing=True)  

when I run the above code, I am getting this warning:

Epoch 1/100  WARNING:tensorflow:Layers in a Sequential model should only have a single input tensor, but we receive a <class 'dict'> input: {'sex': <tf.Tensor 'ExpandDims_4:0' shape=(None, 1) dtype=int64>, 'age': <tf.Tensor 'ExpandDims:0' shape=(None, 1) dtype=int64>, 'bmi': <tf.Tensor 'ExpandDims_1:0' shape=(None, 1) dtype=float64>, 'smoker': <tf.Tensor 'ExpandDims_5:0' shape=(None, 1) dtype=int64>, 'lab1': <tf.Tensor 'ExpandDims_2:0' shape=(None, 1) dtype=int64>, 'lab2': <tf.Tensor 'ExpandDims_3:0' shape=(None, 1) dtype=int64>}  Consider rewriting this model with the Functional API.  WARNING:tensorflow:Layers in a Sequential model should only have a single input tensor, but we receive a <class 'dict'> input: {'sex': <tf.Tensor 'ExpandDims_4:0' shape=(None, 1) dtype=int64>, 'age': <tf.Tensor 'ExpandDims:0' shape=(None, 1) dtype=int64>, 'bmi': <tf.Tensor 'ExpandDims_1:0' shape=(None, 1) dtype=float64>, 'smoker': <tf.Tensor 'ExpandDims_5:0' shape=(None, 1) dtype=int64>, 'lab1': <tf.Tensor 'ExpandDims_2:0' shape=(None, 1) dtype=int64>, 'lab2': <tf.Tensor 'ExpandDims_3:0' shape=(None, 1) dtype=int64>}  Consider rewriting this model with the Functional API.  1/1 [==============================] - ETA: 0s - loss: -22.8739 - accuracy: 0.2500WARNING:tensorflow:Layers in a Sequential model should only have a single input tensor, but we receive a <class 'dict'> input: {'sex': <tf.Tensor 'ExpandDims_4:0' shape=(None, 1) dtype=int64>, 'age': <tf.Tensor 'ExpandDims:0' shape=(None, 1) dtype=int64>, 'bmi': <tf.Tensor 'ExpandDims_1:0' shape=(None, 1) dtype=float64>, 'smoker': <tf.Tensor 'ExpandDims_5:0' shape=(None, 1) dtype=int64>, 'lab1': <tf.Tensor 'ExpandDims_2:0' shape=(None, 1) dtype=int64>, 'lab2': <tf.Tensor 'ExpandDims_3:0' shape=(None, 1) dtype=int64>}  Consider rewriting this model with the Functional API.  

When model fit finished, the accuracy is zero. I know that the data is not valid, bit having an accuracy of zero is also not expected.

https://stackoverflow.com/questions/65860564/why-am-i-getting-warning-error-when-working-with-tensorflow-use-functional-api January 23, 2021 at 10:50PM