提交 c8f06383 编写于 作者: C chengduoZH

update transformer

上级 ccc29237
......@@ -639,7 +639,8 @@ def wrap_decoder(trg_vocab_size,
if weight_sharing:
predict = layers.matmul(
x=dec_output,
y=fluid.get_var(word_emb_param_names[0]),
y=fluid.default_main_program().global_block().var(
word_emb_param_names[0]),
transpose_y=True)
else:
predict = layers.fc(input=dec_output,
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册