Pytorch与tensorflow转换

替换方法：

matrix = tf.get_variable(“Matrix”, [output_size, input_size], dtype=input_.dtype)

matrix = Variable(torch.randn(output_size, input_size))

self.D_l2_loss = tf.constant(0.0)

self.D_l2_loss = torch.Tensor(0.0)

losses = tf.nn.softmax_cross_entropy_with_logits(logits=D_scores, labels=self.D_input_y)

losses = torch.nn.CrossEntropyLoss(D_scores, D_input_y)

tf.reduce_mean(losses)

torch.mean(losses)

embedded_chars = tf.nn.embedding_lookup(W_fe, Feature_input + 1)

embedded_chars = torch.index_select(W_fe, 0, Feature_input + 1)

embedded_chars_expanded = tf.expand_dims(embedded_chars, -1)

embedded_chars_expanded = torch.unsqueeze(embedded_chars, -1)

h_pool = tf.concat(pooled_outputs, 3)

h_pool = torch.cat(pooled_outputs, 3)

tf.transpose(matrix)

matrix.transpose(1, 0)

TensorArray可以看做是具有动态size功能的Tensor数组。通常都是跟while_loop或map_fn结合使用。

10.tf.TensorArray简单教程 https://blog.csdn.net/z2539329562/article/details/80639199

ta.stack(name=None) 将TensorArray中元素叠起来当做一个Tensor输出
ta.unstack(value, name=None) 可以看做是stack的反操作，输入Tensor，输出一个新的TensorArray对象
ta.write(index, value, name=None) 指定index位置写入Tensor
ta.read(index, name=None) 读取指定index位置的Tensor

torch.squeeze
=？？？？？不确定
ta.stack

torch.matmul

tf.matmul

tf.expand_dims

torch.unsqueeze

torch.split( torch.cat([torch.split(input_x, i, 1)[0], self.padding_array], 1),
self.sequence_length, 1)[0]
=有一点不一样
tf.split(tf.concat([tf.split(input_x, [i, self.sequence_length - i], 1)[0], self.padding_array], 1),[self.sequence_length, i], 1)[0]

tf.squeeze()

torch.squeeze

16，
sub_goal = tf.nn.l2_normalize(sub_goal, 1)

sub_goal = sklearn.preprocessing.normalize(sub_goal , norm=‘l2’)

tf.log(tf.nn.softmax(）

F.log_softmax（）

tf.multinomial(log_prob, 1)

torch.multinomial(log_prob, 1, replacement=True)

tf.multiply

*或者torch.mul

tf.one_hot
=？？？
onehot_encoder = OneHotEncoder(sparse=False)
onehot_encoder.fit_transform(next_token)

tf.reduce_sum（a, 0）

a.sum(0)