import tensorflow as tf
x=tf.random.normal([1,3])
w=tf.ones([3,2])
b=tf.ones([2])
y = tf.constant([0, 1])
with tf.GradientTape() as tape:
tape.watch([w, b])
logits = tf.sigmoid(x@w+b)
loss = tf.reduce_mean(tf.losses.MSE(y, logits))
grads = tape.gradient(loss, [w, b])
print('w grad:', grads[0])
print('b grad:', grads[1])
tensorflow2.0基本操作多输出感知器
最新推荐文章于 2022-04-17 13:22:32 发布