Open VV20192019 opened 6 months ago
Ah this is terrible practice by me, I haven't pinned my package versions! Oh no.
Thank you for your issue! I'm not 100% sure which part of the notebook you are referencing. Can you please point me to the exact cell/line?
TFP Version 0.16.0 TF Version 2.8.0
To obtain the correct results, the training data X must input all dimensions, and the author only inputs one dimension, which cannot reproduce the results in the graph
x_train_con = np.concatenate([X[0:200, :], xtrain[0:200][:, None]], axis=-1) x_test_con = np.concatenate([X[200:250, :], xtest[:, None]], axis=-1) model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=0.05), loss=negloglik) model.fit(x_train_con, ytrain, epochs=500, verbose=1)