1. RuntimeError: loss passed to Optimizer.compute_gradients should be a function when eager execution is enabled.

optimizer = tf.compat.v1.train.GradientDescentOptimizer(0.5)
train = optimizer.minimize(loss)

關(guān)于tf.compat.v1.train.GradientDescentOptimizer的介紹參考https://tensorflow.google.cn/api_docs/python/tf/compat/v1/train/GradientDescentOptimizer?hl=en
Methods中minimize

When eager execution is enabled, loss should be a Python function that takes no arguments and computes the value to be minimized.

當(dāng)eager execution開啟的時候，loss應(yīng)該是一個Python函數(shù)。

關(guān)于eager execution的介紹，參考https://tensorflow.google.cn/guide/eager?hl=en

In Tensorflow 2.0, eager execution is enabled by default.
在Tensorflow 2.0 中，eager execution 是默認(rèn)開啟的。

所以，需要先關(guān)閉eager execution

tf.compat.v1.disable_eager_execution()