From f787ebd26e3ae3546604ab0fe6a7f1f589beab50 Mon Sep 17 00:00:00 2001 From: FrostML <380185688@qq.com> Date: Sun, 7 Feb 2021 06:06:56 +0000 Subject: [PATCH] update doc to 2.0 --- .../language_model/transformer-xl/README.md | 22 ++++++++++--------- .../machine_translation/transformer/README.md | 2 +- .../transformers/transformer/modeling.py | 10 +++++---- 3 files changed, 19 insertions(+), 15 deletions(-) diff --git a/examples/language_model/transformer-xl/README.md b/examples/language_model/transformer-xl/README.md index 8c06dc606aa9..c0d52d2b7b0b 100644 --- a/examples/language_model/transformer-xl/README.md +++ b/examples/language_model/transformer-xl/README.md @@ -23,21 +23,23 @@ ### 安装说明 -1. paddle安装 + * PaddlePaddle安装 - 本项目依赖于 PaddlePaddle 2.0rc1及以上版本或适当的develop版本，请参考 [安装指南](https://www.paddlepaddle.org.cn/install/quick) 进行安装 + 本项目依赖于 PaddlePaddle 2.0 及以上版本或适当的develop版本，请参考 [安装指南](https://www.paddlepaddle.org.cn/install/quick) 进行安装 -2. 下载代码 + * PaddleNLP安装 - 克隆代码库到本地 - -3. 环境依赖 +```sh +pip install paddlenlp==2.0.0rc +``` - 该模型使用PaddlePaddle，关于环境依赖部分，请先参考PaddlePaddle[安装说明](https://www.paddlepaddle.org.cn/documentation/docs/zh/install/index_cn.html)关于环境依赖部分的内容。 - 此外，需要另外涉及： - * attrdict - * pyyaml + * 环境依赖 + - attrdict + - pyyaml +``` sh +pip install attrdict pyyaml +``` ### 数据准备 diff --git a/examples/machine_translation/transformer/README.md b/examples/machine_translation/transformer/README.md index dd58ee2fcf07..80d503336d89 100644 --- a/examples/machine_translation/transformer/README.md +++ b/examples/machine_translation/transformer/README.md @@ -47,7 +47,7 @@ Decoder 具有和 Encoder 类似的结构，只是相比于组成 Encoder 的 la * PaddleNLP安装 ```shell -pip install paddlenlp>=2.0.0rc +pip install paddlenlp==2.0.0rc ``` * 环境依赖 diff --git a/paddlenlp/transformers/transformer/modeling.py b/paddlenlp/transformers/transformer/modeling.py index a3d25fffff6c..9125b94d6a16 100644 --- a/paddlenlp/transformers/transformer/modeling.py +++ b/paddlenlp/transformers/transformer/modeling.py @@ -268,7 +268,9 @@ def __init__(self, transpose_y=True) else: self.linear = nn.Linear( - input_dim=d_model, output_dim=trg_vocab_size, bias_attr=False) + in_features=d_model, + out_features=trg_vocab_size, + bias_attr=False) def forward(self, src_word, trg_word): src_max_len = paddle.shape(src_word)[-1] @@ -294,21 +296,21 @@ def forward(self, src_word, trg_word): enc_input = F.dropout( src_emb, p=self.dropout, training=self.training) if self.dropout else src_emb - + trg_emb = self.trg_word_embedding(trg_word) trg_pos_emb = self.trg_pos_embedding(trg_pos) trg_emb = trg_emb + trg_pos_emb dec_input = F.dropout( trg_emb, p=self.dropout, training=self.training) if self.dropout else trg_emb - + dec_output = self.transformer( enc_input, dec_input, src_mask=src_slf_attn_bias, tgt_mask=trg_slf_attn_bias, memory_mask=trg_src_attn_bias) - + predict = self.linear(dec_output) return predict