Tensorflow，从 csv 文件读取数据时出现 Conv1D 错误

Question

I am a newbie at TensorFlow and try to set up the prediction model for graduate school acceptance.

By using pandas, I was able to prepare the training data set as a list.

The data of the CSV file are composed of 4 columns and 425 rows (the first column (admit) is the answer, 3 more columns (gre, gpa, rank) are the training data) as shown below.

import pandas as pd

data = pd.read_csv('gpascore.csv')
data = data.dropna()

data_y = data['admit'].values
data_x = []

for i, rows in data.iterrows():
   data_x.append([rows['gre'], rows['gpa'], rows['rank']])

And I tried to revise a prediction model by adding 1D convolution layer and maxpooling layer as below. Originally, the prediction model includes only Dense layers.

import numpy as np    
import tensorflow as tf

model = tf.keras.models.Sequential([
   tf.keras.layers.Conv1D(32, 3, activation='relu', input_shape=(3)),
   tf.keras.layers.MaxPooling1D(1),
   tf.keras.layers.Flatten(),
   tf.keras.layers.Dense(64, activation='tanh'),   
   tf.keras.layers.Dense(1, activation='sigmoid'),
])

model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
model.fit(np.array(data_x), np.array(data_y), epochs=100)

score = model.evaluate(data_x, data_y)
print(score)

pred = model.predict([ [750, 3.70, 3], [400, 2.2, 1] ])
print(pred)

However, I have some errors as shown below.

And I tried to change the batch size and kernal size, but I still have the same errors.

Could someone help me to correct the errors, please?

Thank you

=================================================================

I really appreciate the help. And I updated the new errors below.

Answer 1

Problem with the COnv1D line

tf.keras.layers.Conv1D(32, 3, activation='relu', input_shape=(3))

Conv1D over sequences expects 3dimension input. [batch, time_steps, single_vector] Add extra dimension to your input. First convert dataframe to numpy_array

data_x = data_x.to_numpy()
data_x = data_x.reshape(-1,425,3)

model = tf.keras.models.Sequential([
   tf.keras.layers.Conv1D(32, 3, activation='relu', input_shape=[None,3]),
   tf.keras.layers.MaxPooling1D(1),
   tf.keras.layers.Flatten(),
   tf.keras.layers.Dense(64, activation='tanh'),   
   tf.keras.layers.Dense(1, activation='sigmoid'),
])

model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
model.fit(X, np.array(data_y), epochs=100)