不幸的是,我不知道该问题的标题,也许有人可以更改?
如何用一种优雅的方式替换以下for循环?
#tensor.shape -> (batchsize,100)
#indices.shape -> (batchsize,100)
liste = []
for i in range(tensor.shape[0]):
liste.append(tf.gather(tensor[i,:], indices[i,:10]))
new_tensor = tf.stack(liste)
答案 0 :(得分:1)
这应该可以解决问题:
new_tensor = tf.gather(tensor, axis=-1, indices=indices[:, :10], batch_dims=1)
这里有一个最小的可复制示例:
import tensorflow as tf
# for version 1.x
#tf.enable_eager_execution()
tensor = tf.random.normal((2, 10))
indices = tf.random.uniform(shape=[2, 10], minval=0, maxval=4, dtype=tf.int32)
liste = []
for i in range(tensor.shape[0]):
liste.append(tf.gather(tensor[i,:], indices[i,:5]))
new_tensor = tf.stack(liste)
print('tensor: ')
print(tensor)
print('new_tensor: ')
print(new_tensor)
new_tensor_v2 = tf.gather(tensor, axis=-1, indices=indices[:, :5], batch_dims=1)
print('new_tensor_v2: ')
print(new_tensor_v2)