From 1925731a45e66d7ab253aed04f01117491b213dc Mon Sep 17 00:00:00 2001
From: 0YuanZhang0 <953963890@qq.com>
Date: Wed, 16 Oct 2019 18:42:02 +0800
Subject: [PATCH] fix_dnet_bug (#3623) (#3625)

---
 .../multi_task_learning/README.md             |  7 ++-
 .../multi_task_learning/scripts/args.py       | 48 +++++++++++++++++++
 .../scripts/convert_model_params.py           |  2 +-
 3 files changed, 55 insertions(+), 2 deletions(-)
 create mode 100644 PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/args.py

diff --git a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/README.md b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/README.md
index efb254e0..948820a6 100644
--- a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/README.md
+++ b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/README.md
@@ -43,8 +43,13 @@ The downloaded data will be saved into `data/mrqa` (combined MRQA training and d
 In our MTL experiments, we use BERT as our shared encoder. The parameters are initialized from the Whole Word Masking BERT (BERTwwm), further fine-tuned on the SQuAD 2.0 task with synthetic generated question answering corpora. The model parameters in Tensorflow format can be downloaded [here](https://worksheets.codalab.org/worksheets/0x3852e60a51d2444680606556d404c657). The following command can be used to convert the parameters to the format that is readable for PaddlePaddle.
 
 ```
-cd scripts && python convert_model_params.py  --init_tf_checkpoint tf_model --fluid_params_dir paddle_model && cd ..
+1、cd scripts
+2、download cased_model_01.tar.gz from link
+3、mkdir cased_model_01 && mv cased_model_01.tar.gz cased_model_01 && cd cased_model_01 && tar -xvf cased_model_01.tar.gz && cd ..
+4、python convert_model_params.py --init_tf_checkpoint cased_model_01/model.ckpt --fluid_params_dir params
+5、mkdir fluid_models && mv cased_model_01/vocab.txt cased_model_01/bert_config.json params fluid_models 
 ```
+
 Alternatively, user can directly **download the parameters that we have converted**: 
 
 ```
diff --git a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/args.py b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/args.py
new file mode 100644
index 00000000..b9be634f
--- /dev/null
+++ b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/args.py
@@ -0,0 +1,48 @@
+#   Copyright (c) 2019 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Arguments for configuration."""
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import six
+import argparse
+
+
+def str2bool(v):
+    # because argparse does not support to parse "true, False" as python
+    # boolean directly
+    return v.lower() in ("true", "t", "1")
+
+
+class ArgumentGroup(object):
+    def __init__(self, parser, title, des):
+        self._group = parser.add_argument_group(title=title, description=des)
+
+    def add_arg(self, name, type, default, help, **kwargs):
+        type = str2bool if type == bool else type
+        self._group.add_argument(
+            "--" + name,
+            default=default,
+            type=type,
+            help=help + ' Default: %(default)s.',
+            **kwargs)
+
+
+def print_arguments(args):
+    print('-----------  Configuration Arguments -----------')
+    for arg, value in sorted(six.iteritems(vars(args))):
+        print('%s: %s' % (arg, value))
+    print('------------------------------------------------')
diff --git a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/convert_model_params.py b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/convert_model_params.py
index fb99bb01..b1c0652c 100644
--- a/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/convert_model_params.py
+++ b/PaddleNLP/Research/MRQA2019-D-NET/multi_task_learning/scripts/convert_model_params.py
@@ -20,7 +20,7 @@ from __future__ import print_function
 import numpy as np
 import argparse
 import collections
-from utils.args import print_arguments
+from args import print_arguments
 import tensorflow as tf
 import paddle.fluid as fluid
 from tensorflow.python import pywrap_tensorflow
-- 
GitLab