从数据框创建TensorFlow模型

I'm new to tensorflow and I was wondering how to make a simple model with tensorflow v1.12 ONLY not using ts.keras for the following dataframe:

   from sklearn.datasets import make_moons
import pandas as pd
X, y = make_moons(n_samples=500, noise=0.1)
xdf=pd.DataFrame(X,columns=['f1','f2'])
ydf=pd.DataFrame(y,columns=['label'])
display(xdf.head(5))
display(ydf.head(5))

enter image description here

I already created a model with ts.keras but I'm just curious how to do it with Tensorflow only