Closed
Description
I have windows 10 CPU with:
Python 3.9.13
tensorflow 2.10.0
I'm testing a simple XOR training:
var x = np.array(new float[,] { { 0, 0 }, { 0, 1 }, { 1, 0 }, { 1, 1 } });
var y = np.array(new float[,] { { 0 }, { 1 }, { 1 }, { 0 } });
var model = keras.Sequential();
model.add(keras.Input(2));
model.add(keras.layers.Dense(32, keras.activations.Relu)); //
model.add(keras.layers.Dense(64, keras.activations.Relu));
model.add(keras.layers.Dense(1, keras.activations.Sigmoid));
model.compile(keras.optimizers.Adam(), keras.losses.MeanSquaredError(), new[] { "accuracy" });
model.fit(x, y, epochs: 800, verbose: 2);
print(model.predict(x, 4));The results from print command are as follow:
tf.Tensor: shape=(4, 1), dtype=float32, numpy=array([[0,5],
[0,492741],
[0,4584209],
[0,4671426]])
PROBLEM
- When I run same sample on python/keras all is OK
- The console does not show the training messages!
- The c# results are incorrect!
What is wrong? Is there a way to get simmilar results for python/tensorflow/keras and the Tensorflow.Net?