Why ais my output "nan", keras model prediction

Question

I'm trying to make an AI attempting to predict numbers from prime number sequence, but my model outputs "[[nan]]". My csv file is formatted like this:

number of the prime, prime

and it contains 78498 lines following this pattern.

I tried looking into the model and it turns out I have input shape of (None, 1).

Does anyone know how to fix this?

here is my code:

import pandas as pd
import tensorflow as tf
from tensorflow import keras

data = pd.read_csv('primes.csv')
data = np.array(data, dtype=float)

data = data.T
numbers = np.array(data[0])
primes = np.array(data[1])

model = tf.keras.Sequential([keras.layers.Dense(units=1, input_shape=[1])])
model.compile(optimizer='sgd', loss='mean_squared_error')
model.fit(numbers, primes, epochs=10)

results = model.predict([1000000])

print(results)

model.save('model.h5')

I use tensorflow: 2.8.0

Answer 1

This is a sample transformation class you can use

The categorical value in this case is status_mapping

statusmapping={
    'OPEN': 0,
    'PENDING': 1,
    'PAYMENT_PENDING': 2,
            }
df['status_mapping'] = df['status'].map(statusmapping)
df

You can use these notebook for NAN values

https://www.kaggle.com/code/parulpandey/a-guide-to-handling-missing-values-in-python