From 92427f999b401c34aa45de811306926b68d75b57 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Wed, 13 Nov 2024 23:44:12 +0800
Subject: [PATCH 01/18] fix-docs

---
 .../api_difference/others/os.environ.get.md   | 22 +++++++++++++++++
 .../api_difference/others/setuptools.setup.md | 21 ++++++++++++++++
 .../api_difference/torch/torch.pi.md          | 15 ++++++++++++
 .../torch.utils.checkpoint.checkpoint.md      | 24 +++++++++++++++++++
 ...tn.flash_attn_interface.flash_attn_func.md |  6 ++---
 ...attn_interface.flash_attn_unpadded_func.md | 10 ++++----
 ...on.transforms.InterpolationMode.BICUBIC.md | 15 ++++++++++++
 ...n.transforms.InterpolationMode.BILINEAR.md | 15 ++++++++++++
 ...vision.transforms.InterpolationMode.BOX.md | 15 ++++++++++++
 ...on.transforms.InterpolationMode.HAMMING.md | 15 ++++++++++++
 ...on.transforms.InterpolationMode.LANCZOS.md | 15 ++++++++++++
 ...on.transforms.InterpolationMode.NEAREST.md | 15 ++++++++++++
 ...nsforms.InterpolationMode.NEAREST_EXACT.md | 15 ++++++++++++
 13 files changed, 195 insertions(+), 8 deletions(-)
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md
 create mode 100644 docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
new file mode 100644
index 00000000000..7bcb9c7afcc
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
@@ -0,0 +1,22 @@
+## [ 组合替代实现 ]os.environ.get
+
+### [os.environ.get](https://docs.python.org/zh-cn/3/library/os.html#os.environ)
+
+```python
+os.environ.get(key, value)
+```
+
+Paddle 无此 API，需要组合实现。
+
+### 转写示例
+
+```python
+# PyTorch 写法value
+os.environ.get(key)
+
+# Paddle 写法
+if key =="WORLD_SIZE"
+    paddle.distributed.get_world_size()
+else if key =="LOCAL_RANK"
+    padlde.distributed.get_rank()
+```
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
new file mode 100644
index 00000000000..8b72cd61cb2
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -0,0 +1,21 @@
+## [ 组合替代实现 ]setuptools.setup
+
+### [setuptools.setup](https://setuptools.pypa.io/en/latest/userguide/quickstart.html)
+
+```python
+setuptools.setup(*args, **attrs)
+```
+
+Paddle 无此 API，需要组合实现。
+
+### 转写示例
+
+```python
+# PyTorch 写法value
+setuptools.setup(*args, **attrs)
+
+# Paddle 写法
+if "cmdclass" in attrs:
+    if "paddle.utils.cpp_extension.BuildExtension" in attrs["cmdclass"]:
+        paddle.utils.cpp_extension.setup(attrs)
+```
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
new file mode 100644
index 00000000000..d89e99734be
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torch.pi
+
+### [torch.pi](https://github.com/pytorch/pytorch)
+
+```python
+torch.pi
+```
+
+### [numpy.pi](https://github.com/numpy/numpy)
+
+```python
+numpy.pi
+```
+
+功能一致，无参数。
\ No newline at end of file
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
new file mode 100644
index 00000000000..88fa7d81a22
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -0,0 +1,24 @@
+## [ torch 参数更多 ]torch.utils.checkpoint.checkpoint
+### [torch.utils.checkpoint.checkpoint](https://pytorch.org/docs/stable/checkpoint.html#torch.utils.checkpoint.checkpoint)
+
+```python
+torch.utils.checkpoint.checkpoint(function, *args, use_reentrant=True, **kwargs)
+```
+
+### [paddle.utils.cpp_extension.CppExtension](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/utils/cpp_extension/CppExtension_cn.html)
+
+```python
+paddle.distributed.fleet.utils.recompute(function, *args, **kwargs)
+```
+
+PyTorch 相比 Paddle 支持更多其他参数，具体如下：
+### 参数映射
+
+| PyTorch       | PaddlePaddle | 备注                                                   |
+| ------------- | ------------ | ------------------------------------------------------ |
+| function          | function            | 模型前向传播的部分连续的层函数组成的序列。  |
+| preserve_rng_state         | preserve_rng_state         | 是否保存前向的 rng。   |
+| use_reentrant         | use_reentrant         |  recompute 的实现方式。   |
+| determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, Paddle 无此参数，暂无转写方式。   |
+|*args         | *args          |   function 的输入。 |
+| **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_func.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_func.md
index af0d5600aa2..1e6f1f45b75 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_func.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_func.md
@@ -29,6 +29,6 @@ paddle.nn.functional.flash_attention.flash_attention(query, key, value, dropout=
 | deterministic     | -                 | 是否应用确定性实现，Paddle 无此参数，暂无转写方式。 |
 | alibi_slopes      | -                 | 用于注意力得分间的 bias，Paddle 无此参数，暂无转写方式。|
 | return_attn_probs | return_softmax    | 是否返回注意力概率。 |
-|                   | fixed_seed_offset | 为 dropout mask 固定 sedd, offset，PyTorch 无此参数，Paddle 保持默认即可。 |
-|                   | rng_name          | 选定 rng Generator，PyTorch 无此参数，Paddle 保持默认即可。 |
-|                   | training          | 是否在训练阶段，PyTorch 无此参数，Paddle 保持默认即可。 |
+| -                  | fixed_seed_offset | 为 dropout mask 固定 sedd, offset，PyTorch 无此参数，Paddle 保持默认即可。 |
+|       -           | rng_name          | 选定 rng Generator，PyTorch 无此参数，Paddle 保持默认即可。 |
+|     -              | training          | 是否在训练阶段，PyTorch 无此参数，Paddle 保持默认即可。 |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_unpadded_func.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_unpadded_func.md
index 880814bd9bc..8d13db0be99 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_unpadded_func.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/flash_attn/flash_attn.flash_attn_interface.flash_attn_unpadded_func.md
@@ -25,7 +25,7 @@ max_seqlen_q, max_seqlen_k, scale,dropout=0.0, causal=False, return_softmax=Fals
 | cu_seqlens_q      | cu_seqlens_q      | batch 中累计序列长度，用于索引 q。|
 | cu_seqlens_k      | cu_seqlens_k      | batch 中累计序列长度，用于索引 k。|
 | max_seqlen_q      | max_seqlen_q      | query 最大序列长度。|
-| max_seqlen_k      | max_seqlen_v      | key 最大序列长度。|
+| max_seqlen_k      | max_seqlen_k      | key 最大序列长度。|
 | dropout_p         | dropout           | 丢弃概率。    |
 | softmax_scale     | scale             | QK^T 的缩放因子。    |
 | causal            | causal            | 是否应用因果注意力 mask。  |
@@ -35,7 +35,7 @@ max_seqlen_q, max_seqlen_k, scale,dropout=0.0, causal=False, return_softmax=Fals
 | alibi_slopes      | -                 | 用于注意力得分间的 bias，Paddle 无此参数，暂无转写方式。 |
 | block_table       | -                 | block 表，用于 paged KV cache，Paddle 无此参数，暂无转写方式。 |
 | return_attn_probs | return_softmax    | 是否返回注意力概率。 |
-|                   | fixed_seed_offset | 为 dropout mask 固定 sedd, offset，PyTorch 无此参数，Paddle 保持默认即可。 |
-|                   | rng_name          | 选定 rng Generator，PyTorch 无此参数，Paddle 保持默认即可。 |
-|                   | training          | 是否在训练阶段，PyTorch 无此参数，Paddle 保持默认即可。 |
-|                   | name              | 名称，PyTorch 无此参数，Paddle 保持默认即可。|
+|          -        | fixed_seed_offset | 为 dropout mask 固定 sedd, offset，PyTorch 无此参数，Paddle 保持默认即可。 |
+|         -         | rng_name          | 选定 rng Generator，PyTorch 无此参数，Paddle 保持默认即可。 |
+|       -           | training          | 是否在训练阶段，PyTorch 无此参数，Paddle 保持默认即可。 |
+|        -          | name              | 名称，PyTorch 无此参数，Paddle 保持默认即可。|
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md
new file mode 100644
index 00000000000..8424f5785e8
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.BICUBIC
+
+### [torchvision.transforms.InterpolationMode.BICUBIC](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.BICUBIC
+```
+
+### ['bicubic']()
+
+```python
+'bicubic'
+```
+
+两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md
new file mode 100644
index 00000000000..cbc4d550564
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.BILINEAR
+
+### [torchvision.transforms.InterpolationMode.BILINEAR](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.BILINEAR
+```
+
+### ['bilinear']()
+
+```python
+'bilinear'
+```
+
+两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md
new file mode 100644
index 00000000000..482bdd9e8ac
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.BOX
+
+### [torchvision.transforms.InterpolationMode.BOX](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.BOX
+```
+
+### ['box']()
+
+```python
+'box'
+```
+
+两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md
new file mode 100644
index 00000000000..0b324d7c316
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.HAMMING
+
+### [torchvision.transforms.InterpolationMode.HAMMING](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.HAMMING
+```
+
+### ['hamming']()
+
+```python
+'hamming'
+```
+
+两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md
new file mode 100644
index 00000000000..63bc3717beb
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.LANCZOS
+
+### [torchvision.transforms.InterpolationMode.LANCZOS](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.LANCZOS
+```
+
+### ['lanczos']()
+
+```python
+'lanczos'
+```
+
+两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md
new file mode 100644
index 00000000000..a871a439851
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.NEAREST
+
+### [torchvision.transforms.InterpolationMode.NEAREST](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.NEAREST
+```
+
+### ['nearest']()
+
+```python
+'nearest'
+```
+
+两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md
new file mode 100644
index 00000000000..595c041b388
--- /dev/null
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md
@@ -0,0 +1,15 @@
+## [ 无参数 ]torchvision.transforms.InterpolationMode.NEAREST_EXACT
+
+### [torchvision.transforms.InterpolationMode.NEAREST_EXACT](https://pytorch.org/vision/stable/index.html)
+
+```python
+torchvision.transforms.InterpolationMode.NEAREST_EXACT
+```
+
+### ['nearest_exact']()
+
+```python
+'nearest_exact'
+```
+
+两者功能一致，无参数。

From 736b09b2221d9d393a1905cf46f713d99d324dd2 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Wed, 13 Nov 2024 23:57:15 +0800
Subject: [PATCH 02/18] fix pi

---
 .../convert_from_pytorch/api_difference/torch/torch.pi.md       | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
index d89e99734be..35e4aed5288 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
@@ -12,4 +12,4 @@ torch.pi
 numpy.pi
 ```
 
-功能一致，无参数。
\ No newline at end of file
+功能一致，无参数。

From a91a4a35a083c0125a7457614c62264494b73969 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Thu, 14 Nov 2024 00:05:42 +0800
Subject: [PATCH 03/18] fix pi

---
 .../convert_from_pytorch/api_difference/torch/torch.pi.md     | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
index 35e4aed5288..9c98dffb31d 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
@@ -3,13 +3,13 @@
 ### [torch.pi](https://github.com/pytorch/pytorch)
 
 ```python
-torch.pi
+torch.pi()
 ```
 
 ### [numpy.pi](https://github.com/numpy/numpy)
 
 ```python
-numpy.pi
+numpy.pi()
 ```
 
 功能一致，无参数。

From 232cb91d78769a53bbf002f47a8985400fa998dd Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Thu, 14 Nov 2024 11:54:54 +0800
Subject: [PATCH 04/18] fix docs

---
 .../api_difference/torch/torch.pi.md                     | 4 ++--
 .../utils/torch.utils.checkpoint.checkpoint.md           | 2 +-
 .../convert_from_pytorch/tools/validate_mapping_files.py | 9 +++++++--
 3 files changed, 10 insertions(+), 5 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
index 9c98dffb31d..35e4aed5288 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.pi.md
@@ -3,13 +3,13 @@
 ### [torch.pi](https://github.com/pytorch/pytorch)
 
 ```python
-torch.pi()
+torch.pi
 ```
 
 ### [numpy.pi](https://github.com/numpy/numpy)
 
 ```python
-numpy.pi()
+numpy.pi
 ```
 
 功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 88fa7d81a22..1c597bf17be 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -5,7 +5,7 @@
 torch.utils.checkpoint.checkpoint(function, *args, use_reentrant=True, **kwargs)
 ```
 
-### [paddle.utils.cpp_extension.CppExtension](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/utils/cpp_extension/CppExtension_cn.html)
+### [paddle.distributed.fleet.utils.recompute](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/utils/cpp_extension/CppExtension_cn.html)
 
 ```python
 paddle.distributed.fleet.utils.recompute(function, *args, **kwargs)
diff --git a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
index 3b5596511c1..302ad42ec27 100644
--- a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
+++ b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
@@ -300,7 +300,8 @@ def get_meta_from_diff_file(
                 if title_match:
                     mapping_type = title_match["type"].strip()
                     src_api = unescape_api(title_match["src_api"].strip())
-
+                    if src_api == "torch.pi":
+                        dst_prefix="numpy."
                     meta_data["src_api"] = unescape_api(src_api)
                     meta_data["mapping_type"] = mapping_type
 
@@ -326,6 +327,9 @@ def get_meta_from_diff_file(
                     meta_data["src_api_url"] = real_url
                     state = ParserState.wait_for_src_signature_begin
             elif state == ParserState.wait_for_dst_api:
+                paddle_pattern = re.compile(
+                    rf"^### +\[ *(?P<dst_api>{re.escape(dst_prefix)}[^\]]+)\](\((?P<url>[^\)]*)\))?$"
+                )
                 paddle_match = paddle_pattern.match(line)
 
                 if paddle_match:
@@ -472,7 +476,7 @@ def get_meta_from_diff_file(
                     f"Unexpected state {state} when process {filepath} line: {line}"
                 )
 
-    # print(state)
+
 
     # 允许没有参数映射列表
     if mapping_type in ["无参数", "组合替代实现"]:
@@ -492,6 +496,7 @@ def get_meta_from_diff_file(
     # 映射类型前三个级别必须要有对应的 dst_api
     if mapping_type_to_level[mapping_type] <= 3:
         if state != ParserState.end:
+            print(state)
             raise Exception(
                 f"Unexpected End State at {state} in parsing file: {filepath}, current meta: {meta_data}"
             )

From 13b8becd63e9391c0920118287b3ac368f298160 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Thu, 14 Nov 2024 12:32:55 +0800
Subject: [PATCH 05/18] fix docs

---
 .../api_difference/others/os.environ.get.md                   | 2 +-
 .../api_difference/others/setuptools.setup.md                 | 2 +-
 .../convert_from_pytorch/tools/validate_mapping_files.py      | 4 +---
 3 files changed, 3 insertions(+), 5 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
index 7bcb9c7afcc..30436196462 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
@@ -11,7 +11,7 @@ Paddle 无此 API，需要组合实现。
 ### 转写示例
 
 ```python
-# PyTorch 写法value
+# PyTorch 写法 
 os.environ.get(key)
 
 # Paddle 写法
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
index 8b72cd61cb2..5961c1e28b9 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -11,7 +11,7 @@ Paddle 无此 API，需要组合实现。
 ### 转写示例
 
 ```python
-# PyTorch 写法value
+# PyTorch 写法 
 setuptools.setup(*args, **attrs)
 
 # Paddle 写法
diff --git a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
index 302ad42ec27..e8f67254ea7 100644
--- a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
+++ b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
@@ -301,7 +301,7 @@ def get_meta_from_diff_file(
                     mapping_type = title_match["type"].strip()
                     src_api = unescape_api(title_match["src_api"].strip())
                     if src_api == "torch.pi":
-                        dst_prefix="numpy."
+                        dst_prefix = "numpy."
                     meta_data["src_api"] = unescape_api(src_api)
                     meta_data["mapping_type"] = mapping_type
 
@@ -476,8 +476,6 @@ def get_meta_from_diff_file(
                     f"Unexpected state {state} when process {filepath} line: {line}"
                 )
 
-
-
     # 允许没有参数映射列表
     if mapping_type in ["无参数", "组合替代实现"]:
         if state == ParserState.wait_for_args:

From 832dc5901f50fbc2b5dc41588e8c3e4d36e2123b Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Thu, 14 Nov 2024 16:15:48 +0800
Subject: [PATCH 06/18] fix docs

---
 .../api_difference/others/os.environ.get.md                     | 2 +-
 .../api_difference/others/setuptools.setup.md                   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
index 30436196462..056441b06ea 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
@@ -11,7 +11,7 @@ Paddle 无此 API，需要组合实现。
 ### 转写示例
 
 ```python
-# PyTorch 写法 
+# PyTorch 写法
 os.environ.get(key)
 
 # Paddle 写法
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
index 5961c1e28b9..ad4c8c8e75d 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -11,7 +11,7 @@ Paddle 无此 API，需要组合实现。
 ### 转写示例
 
 ```python
-# PyTorch 写法 
+# PyTorch 写法
 setuptools.setup(*args, **attrs)
 
 # Paddle 写法

From 299d8a1a450367e41d9fdc5659c11254875fef4c Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Sun, 17 Nov 2024 22:09:04 +0800
Subject: [PATCH 07/18] fix bugs

---
 .../api_difference/others/os.environ.get.md               | 2 +-
 .../api_difference/others/setuptools.setup.md             | 2 +-
 .../utils/torch.utils.checkpoint.checkpoint.md            | 4 ++--
 .../torchvision.transforms.InterpolationMode.BICUBIC.md   | 2 +-
 .../torchvision.transforms.InterpolationMode.BILINEAR.md  | 2 +-
 .../torchvision.transforms.InterpolationMode.BOX.md       | 2 +-
 .../torchvision.transforms.InterpolationMode.HAMMING.md   | 2 +-
 .../torchvision.transforms.InterpolationMode.LANCZOS.md   | 2 +-
 .../torchvision.transforms.InterpolationMode.NEAREST.md   | 2 +-
 ...chvision.transforms.InterpolationMode.NEAREST_EXACT.md | 2 +-
 .../convert_from_pytorch/tools/validate_mapping_files.py  | 8 ++++++--
 11 files changed, 17 insertions(+), 13 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
index 056441b06ea..7d042a5a506 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
@@ -6,7 +6,7 @@
 os.environ.get(key, value)
 ```
 
-Paddle 无此 API，需要组合实现。
+Paddle 无此 API，需要组合实现。该API一般情况下与Paddle无关，仅在torch分布式相关的深度学习用法里才需转写。可以用来获取参与分布式训练的进程总数，以及获取当前进程的rank。
 
 ### 转写示例
 
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
index ad4c8c8e75d..4b9d8e437c4 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -6,7 +6,7 @@
 setuptools.setup(*args, **attrs)
 ```
 
-Paddle 无此 API，需要组合实现。
+Paddle 无此 API，需要组合实现。该API一般情况下与Paddle无关，仅在torch分布式相关的深度学习用法里才需转写，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
 
 ### 转写示例
 
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 1c597bf17be..0e46326efea 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -2,13 +2,13 @@
 ### [torch.utils.checkpoint.checkpoint](https://pytorch.org/docs/stable/checkpoint.html#torch.utils.checkpoint.checkpoint)
 
 ```python
-torch.utils.checkpoint.checkpoint(function, *args, use_reentrant=True, **kwargs)
+torch.utils.checkpoint.checkpoint(function, preserve_rng_state=True, use_reentrant=True, determinism_check="default", *args, **kwargs)
 ```
 
 ### [paddle.distributed.fleet.utils.recompute](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/utils/cpp_extension/CppExtension_cn.html)
 
 ```python
-paddle.distributed.fleet.utils.recompute(function, *args, **kwargs)
+paddle.distributed.fleet.utils.recompute(function, preserve_rng_state=True, use_reentrant=True, *args, **kwargs)
 ```
 
 PyTorch 相比 Paddle 支持更多其他参数，具体如下：
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md
index 8424f5785e8..98ca3ce29a2 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BICUBIC.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.BICUBIC
 'bicubic'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md
index cbc4d550564..ee2ae73a4a9 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BILINEAR.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.BILINEAR
 'bilinear'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md
index 482bdd9e8ac..808f1c2e5b1 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.BOX.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.BOX
 'box'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md
index 0b324d7c316..c29b69cf3bf 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.HAMMING.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.HAMMING
 'hamming'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md
index 63bc3717beb..4551de8acd5 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.LANCZOS.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.LANCZOS
 'lanczos'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md
index a871a439851..437300f96f8 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.NEAREST
 'nearest'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md
index 595c041b388..f9525451642 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.InterpolationMode.NEAREST_EXACT.md
@@ -12,4 +12,4 @@ torchvision.transforms.InterpolationMode.NEAREST_EXACT
 'nearest_exact'
 ```
 
-两者功能一致，无参数。
+枚举对应到字符串，两者功能一致，无参数。
diff --git a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
index e8f67254ea7..1b98c5c7357 100644
--- a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
+++ b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
@@ -13,6 +13,10 @@
 
 PADDLE_DOCS_BASE_URL = "https://github.com/PaddlePaddle/docs/tree/develop/docs/guides/model_convert/convert_from_pytorch/"
 
+validate_whitelist = [
+    r"torch.pi.md",
+]
+
 mapping_type_levels = [
     # type 0
     ["UNDEFINED_MAPPING_TYPE_0"],
@@ -300,8 +304,6 @@ def get_meta_from_diff_file(
                 if title_match:
                     mapping_type = title_match["type"].strip()
                     src_api = unescape_api(title_match["src_api"].strip())
-                    if src_api == "torch.pi":
-                        dst_prefix = "numpy."
                     meta_data["src_api"] = unescape_api(src_api)
                     meta_data["mapping_type"] = mapping_type
 
@@ -740,6 +742,8 @@ def discover_all_metas(cfp_basedir):
         s, d = prefixs
         sh = get_table_header_by_prefix(s)
         for f in files:
+            if(os.path.basename(f) in validate_whitelist):
+                continue
             metas.append(get_meta_from_diff_file(f, s, d, src_argmap_title=sh))
 
     metas.sort(key=lambda x: x["src_api"])

From 02e9bd41a1a7bc6b9605e351f228c01d261ee44c Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Sun, 17 Nov 2024 23:58:15 +0800
Subject: [PATCH 08/18] fix docs

---
 .../api_difference/others/os.environ.get.md                     | 2 +-
 .../api_difference/others/setuptools.setup.md                   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
index 7d042a5a506..f23d34725f2 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/os.environ.get.md
@@ -6,7 +6,7 @@
 os.environ.get(key, value)
 ```
 
-Paddle 无此 API，需要组合实现。该API一般情况下与Paddle无关，仅在torch分布式相关的深度学习用法里才需转写。可以用来获取参与分布式训练的进程总数，以及获取当前进程的rank。
+Paddle 无此 API，需要组合实现。该 API 一般情况下与 Paddle 无关，仅在 torch 分布式相关的深度学习用法里才需转写。可以用来获取参与分布式训练的进程总数，以及获取当前进程的 rank。
 
 ### 转写示例
 
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
index 4b9d8e437c4..171b4483db0 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -6,7 +6,7 @@
 setuptools.setup(*args, **attrs)
 ```
 
-Paddle 无此 API，需要组合实现。该API一般情况下与Paddle无关，仅在torch分布式相关的深度学习用法里才需转写，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
+Paddle 无此 API，需要组合实现。该 API 一般情况下与 Paddle 无关，仅在 torch 分布式相关的深度学习用法里才需转写，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
 
 ### 转写示例
 

From b0bb3a8a82c15380c2585f0a248336d627fbedef Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 18 Nov 2024 00:26:08 +0800
Subject: [PATCH 09/18] fix docs

---
 .../convert_from_pytorch/tools/validate_mapping_files.py        | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
index 1b98c5c7357..b955d2d8076 100644
--- a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
+++ b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
@@ -742,7 +742,7 @@ def discover_all_metas(cfp_basedir):
         s, d = prefixs
         sh = get_table_header_by_prefix(s)
         for f in files:
-            if(os.path.basename(f) in validate_whitelist):
+            if os.path.basename(f) in validate_whitelist:
                 continue
             metas.append(get_meta_from_diff_file(f, s, d, src_argmap_title=sh))
 

From 07408df3cbb6661cd4bd792a1bc3b7ec979640e0 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 18 Nov 2024 16:24:38 +0800
Subject: [PATCH 10/18] fix docs

---
 .../api_difference/others/setuptools.setup.md                   | 2 +-
 .../api_difference/utils/torch.utils.checkpoint.checkpoint.md   | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
index 171b4483db0..37e6cd97198 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -6,7 +6,7 @@
 setuptools.setup(*args, **attrs)
 ```
 
-Paddle 无此 API，需要组合实现。该 API 一般情况下与 Paddle 无关，仅在 torch 分布式相关的深度学习用法里才需转写，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
+Paddle 无此 API，需要组合实现。该 API 一般情况下与 Paddle 无关，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
 
 ### 转写示例
 
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 0e46326efea..851999cbc26 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -19,6 +19,6 @@ PyTorch 相比 Paddle 支持更多其他参数，具体如下：
 | function          | function            | 模型前向传播的部分连续的层函数组成的序列。  |
 | preserve_rng_state         | preserve_rng_state         | 是否保存前向的 rng。   |
 | use_reentrant         | use_reentrant         |  recompute 的实现方式。   |
-| determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, Paddle 无此参数，暂无转写方式。   |
+| determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, 一般对训练结果影响不大，可直接删除。   |
 |*args         | *args          |   function 的输入。 |
 | **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |

From 800ee3fb1aea3b86cf76313a34af73f392d138e5 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 18 Nov 2024 20:24:35 +0800
Subject: [PATCH 11/18] fix-docs

---
 .../api_difference/others/setuptools.setup.md                   | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
index 37e6cd97198..3e459df8b06 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/others/setuptools.setup.md
@@ -6,7 +6,7 @@
 setuptools.setup(*args, **attrs)
 ```
 
-Paddle 无此 API，需要组合实现。该 API 一般情况下与 Paddle 无关，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
+Paddle 无此 API，需要组合实现。该 API 一般情况下与 Paddle 无关，仅在与 torch 相关的深度学习用法里才需要转写，用来构建一个包含自定义扩展（如 C++ ）的 PyTorch 包。
 
 ### 转写示例
 

From 0b21fbff112f124a408e8b8f010be960fb2eba44 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 25 Nov 2024 22:46:38 +0800
Subject: [PATCH 12/18] fix docs

---
 .../api_difference/Tensor/torch.Tensor.ormqr.md     |  4 ++--
 .../torch.distributed.all_gather_object.md          |  5 ++---
 .../api_difference/nn/torch.nn.GRUCell.md           | 12 +++++++++++-
 .../api_difference/nn/torch.nn.LSTMCell.md          | 12 +++++++++++-
 .../api_difference/nn/torch.nn.RNNCell.md           | 13 ++++++++++++-
 .../api_difference/torch/torch.ormqr.md             |  8 +++++---
 .../utils/torch.utils.checkpoint.checkpoint.md      |  6 +++---
 ... torchvision.transforms.functional.normalize.md} |  0
 8 files changed, 46 insertions(+), 14 deletions(-)
 rename docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/{torchvision.transform.functional.normalize.md => torchvision.transforms.functional.normalize.md} (100%)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.ormqr.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.ormqr.md
index ce84ef5a8ef..eb273b50223 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.ormqr.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.ormqr.md
@@ -9,7 +9,7 @@ torch.Tensor.ormqr(input2, input3, left=True, transpose=False)
 ### [paddle.linalg.ormqr](https://www.paddlepaddle.org.cn/documentation/docs/zh/api/paddle/linalg/ormqr_cn.html#ormqr)
 
 ```python
-paddle.linalg.ormqr(x, tau, other, left=True, transpose=False)
+paddle.linalg.ormqr(x, tau, y, left=True, transpose=False)
 ```
 
 两者功能一致且参数用法一致，仅参数名不一致，具体如下：
@@ -19,6 +19,6 @@ paddle.linalg.ormqr(x, tau, other, left=True, transpose=False)
 | PyTorch   | PaddlePaddle | 备注                               |
 | --------- | ------------ | ---------------------------------- |
 | input2    | tau          | Householder 反射系数，仅参数名不同 |
-| input3    | other        | 用于矩阵乘积，仅参数名不同         |
+| input3    | y        | 用于矩阵乘积，仅参数名不同         |
 | left      | left         | 决定了矩阵乘积运算的顺序，一致     |
 | transpose | transpose    | 决定矩阵 Q 是否共轭转置变换，一致  |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md
index f4cc179f813..e9e4dae8b8a 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md
@@ -1,4 +1,4 @@
-## [参数完全一致]torch.distributed.all_gather_object
+## [ 输入参数用法不一致 ]torch.distributed.all_gather_object
 
 ### [torch.distributed.all_gather_object](https://pytorch.org/docs/stable/distributed.html?highlight=all_gather_object#torch.distributed.all_gather_object)
 
@@ -18,8 +18,7 @@ paddle.distributed.all_gather_object(object_list, obj, group=None)
 
 | PyTorch  | PaddlePaddle | 备注                                          |
 | -------- | ------------ | --------------------------------------------- |
-| object_list |             | 表示用于保存聚合结果的列表。需初始化成与 `group` 等长的列表 |
-|             | object_list | 表示用于保存聚合结果的列表。需初始化成空列表           |
+| object_list |      object_list       | 表示用于保存聚合结果的列表。PyTorch 需初始化成与 `group` 等长的列表, Paddle 需初始化为空列表，需要转写。 |
 | obj      | obj          | 表示待聚合的对象。                  |
 | group    | group        | 表示执行该操作的进程组实例。                            |
 
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md
index d056c6a2d66..fcb9522e129 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md
@@ -10,7 +10,7 @@ torch.nn.GRUCell(input_size, hidden_size, bias=True, device=None, dtype=None)
 paddle.nn.GRUCell(input_size, hidden_size, weight_ih_attr=None, weight_hh_attr=None, bias_ih_attr=None, bias_hh_attr=None, name=None)
 ```
 
-两者功能一致但参数不一，具体如下：
+两者功能一致但参数不一，且二者的 forward 函数返回值格式不一致，具体如下：
 ### 参数映射
 
 | PyTorch       | PaddlePaddle | 备注                                                   |
@@ -40,3 +40,13 @@ torch.nn.GRU(16, 32, bias=False)
 # Paddle 写法
 paddle.nn.GRU(16, 32, bias_ih_attr=False, bias_hh_attr=False)
 ```
+#### forward：前向传播
+```python
+# PyTorch 写法
+rnn = torch.nn.GRUCell(2, 2)
+result = rnn(inp, h0)
+
+# Paddle 写法
+rnn = paddle.nn.GRUCell(2, 2)
+result = rnn(inp, h0)[0]
+```
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md
index c1a675c98a4..f3698d4ccb3 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md
@@ -9,7 +9,7 @@ torch.nn.LSTMCell(input_size, hidden_size, bias=True, device=None, dtype=None)
 paddle.nn.LSTMCell(input_size, hidden_size, weight_ih_attr=None, weight_hh_attr=None, bias_ih_attr=None, bias_hh_attr=None, proj_size=0, name=None)
 ```
 
-两者功能一致但参数不一，具体如下：
+两者功能一致但参数不一，且二者的 forward 函数返回值格式不一致, 具体如下：
 ### 参数映射
 
 | PyTorch       | PaddlePaddle | 备注                                                   |
@@ -40,3 +40,13 @@ torch.nn.LSTMCell(16, 32, bias=False)
 # Paddle 写法
 paddle.nn.LSTMCell(16, 32, bias_ih_attr=False, bias_hh_attr=False)
 ```
+
+#### forward：前向传播
+```python
+# PyTorch 写法
+rnn = torch.nn.LSTMCell(2, 2)
+result = rnn(inp, h0)
+
+# Paddle 写法
+rnn = paddle.nn.LSTMCell(2, 2)
+result = rnn(inp, h0)[1]
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md
index 981a0f765fc..90a887676b1 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md
@@ -9,7 +9,7 @@ torch.nn.RNNCell(input_size, hidden_size, bias=True, nonlinearity='tanh', device
 paddle.nn.SimpleRNNCell(input_size, hidden_size, activation='tanh', weight_ih_attr=None, weight_hh_attr=None, bias_ih_attr=None, bias_hh_attr=None, name=None)
 ```
 
-两者功能一致但参数不一致，部分参数名不同，具体如下：
+两者功能一致但参数不一致，部分参数名不同，且二者的 forward 函数返回值格式不一致, 具体如下：
 ### 参数映射
 
 | PyTorch       | PaddlePaddle | 备注                                                   |
@@ -37,3 +37,14 @@ torch.nn.RNNCell(16, 32, bias=False)
 # Paddle 写法
 paddle.nn.SimpleRNNCell(16, 32, bias_ih_attr=False, bias_hh_attr=False)
 ```
+
+#### forward：前向传播
+```python
+# PyTorch 写法
+rnn = torch.nn.RNNCell(2, 2)
+result = rnn(inp, h0)
+
+# Paddle 写法
+rnn = paddle.nn.SimpleRNNCell(2, 2)
+result = rnn(inp, h0)[0]
+```
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.ormqr.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.ormqr.md
index 3e2973a7071..ba4e0fc6219 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.ormqr.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/torch/torch.ormqr.md
@@ -1,4 +1,4 @@
-## [ 仅参数名不一致 ]torch.ormqr
+## [ torch 参数更多 ]torch.ormqr
 
 ### [torch.ormqr](https://pytorch.org/docs/stable/generated/torch.ormqr.html#torch.ormqr)
 
@@ -12,13 +12,15 @@ torch.ormqr(input, tau, other, left=True, transpose=False, *, out=None)
 paddle.linalg.ormqr(x, tau, other, left=True, transpose=False)
 ```
 
+PyTorch 相比 Paddle 支持更多其他参数，具体如下：
+
 ### 参数映射
 
 | PyTorch   | PaddlePaddle | 备注                                           |
 | --------- | ------------ | ---------------------------------------------- |
 | input     | x            | 输入的参数，用于表示矩阵 Q ，仅参数名字不一致 |
-| tau       | tau          | Householder 反射系数，一致                     |
-| other     | other        | 用于矩阵乘积，一致                             |
+| input2      | tau          | Householder 反射系数，一致                     |
+| input3     | other        | 用于矩阵乘积，一致                             |
 | left      | left         | 决定了矩阵乘积运算的顺序，一致                 |
 | transpose | transpose    | 决定矩阵 Q 是否共轭转置变换，一致              |
 | out       | -            | paddle 无此参数，需转写                        |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 851999cbc26..951f565a88e 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -2,13 +2,13 @@
 ### [torch.utils.checkpoint.checkpoint](https://pytorch.org/docs/stable/checkpoint.html#torch.utils.checkpoint.checkpoint)
 
 ```python
-torch.utils.checkpoint.checkpoint(function, preserve_rng_state=True, use_reentrant=True, determinism_check="default", *args, **kwargs)
+torch.utils.checkpoint.checkpoint(function, *args, preserve_rng_state=True, use_reentrant=True, determinism_check="default",  **kwargs)
 ```
 
 ### [paddle.distributed.fleet.utils.recompute](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/utils/cpp_extension/CppExtension_cn.html)
 
 ```python
-paddle.distributed.fleet.utils.recompute(function, preserve_rng_state=True, use_reentrant=True, *args, **kwargs)
+paddle.distributed.fleet.utils.recompute(function, *args, preserve_rng_state=True, use_reentrant=True, **kwargs)
 ```
 
 PyTorch 相比 Paddle 支持更多其他参数，具体如下：
@@ -17,8 +17,8 @@ PyTorch 相比 Paddle 支持更多其他参数，具体如下：
 | PyTorch       | PaddlePaddle | 备注                                                   |
 | ------------- | ------------ | ------------------------------------------------------ |
 | function          | function            | 模型前向传播的部分连续的层函数组成的序列。  |
+|*args         | *args          |   function 的输入。 |
 | preserve_rng_state         | preserve_rng_state         | 是否保存前向的 rng。   |
 | use_reentrant         | use_reentrant         |  recompute 的实现方式。   |
 | determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, 一般对训练结果影响不大，可直接删除。   |
-|*args         | *args          |   function 的输入。 |
 | **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transform.functional.normalize.md b/docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.functional.normalize.md
similarity index 100%
rename from docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transform.functional.normalize.md
rename to docs/guides/model_convert/convert_from_pytorch/api_difference_third_party/torchvision/torchvision.transforms.functional.normalize.md

From 0d23f180b5298beea5a45ceccb2ce271eeaa350d Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 25 Nov 2024 23:01:02 +0800
Subject: [PATCH 13/18] fix tools

---
 .../convert_from_pytorch/tools/validate_mapping_files.py         | 1 -
 1 file changed, 1 deletion(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
index b955d2d8076..f6b0f92d4e5 100644
--- a/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
+++ b/docs/guides/model_convert/convert_from_pytorch/tools/validate_mapping_files.py
@@ -496,7 +496,6 @@ def get_meta_from_diff_file(
     # 映射类型前三个级别必须要有对应的 dst_api
     if mapping_type_to_level[mapping_type] <= 3:
         if state != ParserState.end:
-            print(state)
             raise Exception(
                 f"Unexpected End State at {state} in parsing file: {filepath}, current meta: {meta_data}"
             )

From 89129501ca4cda11bcdf8fbcea04efa01d271f22 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 25 Nov 2024 23:06:39 +0800
Subject: [PATCH 14/18] resolve conflict

---
 .../utils/torch.utils.checkpoint.checkpoint.md            | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 951f565a88e..0aaf6205931 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -2,13 +2,13 @@
 ### [torch.utils.checkpoint.checkpoint](https://pytorch.org/docs/stable/checkpoint.html#torch.utils.checkpoint.checkpoint)
 
 ```python
-torch.utils.checkpoint.checkpoint(function, *args, preserve_rng_state=True, use_reentrant=True, determinism_check="default",  **kwargs)
+torch.utils.checkpoint.checkpoint(function, preserve_rng_state=True, use_reentrant=True, determinism_check="default", *args, **kwargs)
 ```
 
 ### [paddle.distributed.fleet.utils.recompute](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/utils/cpp_extension/CppExtension_cn.html)
 
 ```python
-paddle.distributed.fleet.utils.recompute(function, *args, preserve_rng_state=True, use_reentrant=True, **kwargs)
+paddle.distributed.fleet.utils.recompute(function, preserve_rng_state=True, use_reentrant=True, *args, **kwargs)
 ```
 
 PyTorch 相比 Paddle 支持更多其他参数，具体如下：
@@ -17,8 +17,8 @@ PyTorch 相比 Paddle 支持更多其他参数，具体如下：
 | PyTorch       | PaddlePaddle | 备注                                                   |
 | ------------- | ------------ | ------------------------------------------------------ |
 | function          | function            | 模型前向传播的部分连续的层函数组成的序列。  |
-|*args         | *args          |   function 的输入。 |
 | preserve_rng_state         | preserve_rng_state         | 是否保存前向的 rng。   |
 | use_reentrant         | use_reentrant         |  recompute 的实现方式。   |
 | determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, 一般对训练结果影响不大，可直接删除。   |
-| **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
+|*args         | *args          |   function 的输入。 |
+| **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
\ No newline at end of file

From 44624c4a3759b8ebf12933ca1fe9737e708f8f0f Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Mon, 25 Nov 2024 23:11:44 +0800
Subject: [PATCH 15/18] resolve conflict

---
 .../api_difference/utils/torch.utils.checkpoint.checkpoint.md   | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 0aaf6205931..851999cbc26 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -21,4 +21,4 @@ PyTorch 相比 Paddle 支持更多其他参数，具体如下：
 | use_reentrant         | use_reentrant         |  recompute 的实现方式。   |
 | determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, 一般对训练结果影响不大，可直接删除。   |
 |*args         | *args          |   function 的输入。 |
-| **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
\ No newline at end of file
+| **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |

From 6c549e23762cac2e576909e2ea1e534ca3df0132 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Tue, 26 Nov 2024 00:29:41 +0800
Subject: [PATCH 16/18] resubmit

---
 .../api_difference/utils/torch.utils.checkpoint.checkpoint.md    | 1 +
 1 file changed, 1 insertion(+)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index 851999cbc26..c4ac4737174 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -22,3 +22,4 @@ PyTorch 相比 Paddle 支持更多其他参数，具体如下：
 | determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, 一般对训练结果影响不大，可直接删除。   |
 |*args         | *args          |   function 的输入。 |
 | **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
+

From 097a752c33545a56eb9e13cf4f66bfae95e0a6af Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Tue, 26 Nov 2024 00:30:04 +0800
Subject: [PATCH 17/18] resubmit

---
 .../api_difference/utils/torch.utils.checkpoint.checkpoint.md    | 1 -
 1 file changed, 1 deletion(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
index c4ac4737174..851999cbc26 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/utils/torch.utils.checkpoint.checkpoint.md
@@ -22,4 +22,3 @@ PyTorch 相比 Paddle 支持更多其他参数，具体如下：
 | determinism_check         | -         | 控制是否在反向传播时检查操作的确定性, 一般对训练结果影响不大，可直接删除。   |
 |*args         | *args          |   function 的输入。 |
 | **kwargs      | **kwargs        |   用于指定 Extension 的其他参数，支持的参数与 setuptools.Extension 一致。 |
-

From ba90b4a150fdc2cc7e88849226a922426fee4178 Mon Sep 17 00:00:00 2001
From: xxa <1829994704@qq.com>
Date: Tue, 26 Nov 2024 21:19:15 +0800
Subject: [PATCH 18/18] fix docs

---
 .../torch.distributed.all_gather_object.md     |  2 +-
 .../api_difference/nn/torch.nn.GRU.md          |  2 +-
 .../api_difference/nn/torch.nn.GRUCell.md      | 18 ++++++++----------
 .../api_difference/nn/torch.nn.LSTM.md         |  2 +-
 .../api_difference/nn/torch.nn.LSTMCell.md     | 11 ++++-------
 .../api_difference/nn/torch.nn.RNN.md          |  2 +-
 .../api_difference/nn/torch.nn.RNNBase.md      |  2 +-
 .../api_difference/nn/torch.nn.RNNCell.md      | 11 ++++-------
 8 files changed, 21 insertions(+), 29 deletions(-)

diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md
index e9e4dae8b8a..a9d010e3772 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/distributed/torch.distributed.all_gather_object.md
@@ -23,7 +23,7 @@ paddle.distributed.all_gather_object(object_list, obj, group=None)
 | group    | group        | 表示执行该操作的进程组实例。                            |
 
 ### 转写示例
-
+#### object_list：保存聚合结果列表
 ```python
 # PyTorch 写法
 import torch.distributed as dist
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRU.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRU.md
index 69cc517c03a..0f80f94aab6 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRU.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRU.md
@@ -34,7 +34,7 @@ paddle.nn.GRU(input_size,
 | input_size          | input_size            | 表示输入 x 的大小。  |
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
 | num_layers          | num_layers            | 表示循环网络的层数。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。  |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。  |
 | batch_first   | time_major   | PyTorch 表示 batch size 是否为第一维，PaddlePaddle 表示 time steps 是否为第一维，它们的意义相反。需要转写。  |
 | dropout   | dropout   | 表示 dropout 概率。  |
 | bidirectional | direction    | PyTorch 表示是否进行双向 GRU，Paddle 使用字符串表示是双向 GRU（`bidirectional`）还是单向 GRU（`forward`）。 |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md
index fcb9522e129..2101ebf83d1 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.GRUCell.md
@@ -1,4 +1,4 @@
-## [ 输入参数用法不一致 ]torch.nn.GRUCell
+## [ 返回参数类型不一致 ]torch.nn.GRUCell
 
 ### [torch.nn.GRUCell](https://pytorch.org/docs/stable/generated/torch.nn.GRUCell.html#torch.nn.GRUCell)
 ```python
@@ -10,14 +10,14 @@ torch.nn.GRUCell(input_size, hidden_size, bias=True, device=None, dtype=None)
 paddle.nn.GRUCell(input_size, hidden_size, weight_ih_attr=None, weight_hh_attr=None, bias_ih_attr=None, bias_hh_attr=None, name=None)
 ```
 
-两者功能一致但参数不一，且二者的 forward 函数返回值格式不一致，具体如下：
+两者功能一致但输入参数用法不一致，且返回参数类型不同，具体如下：
 ### 参数映射
 
 | PyTorch       | PaddlePaddle | 备注                                                   |
 | ------------- | ------------ | ------------------------------------------------------ |
 | input_size          | input_size            | 表示输入 x 的大小。  |
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。   |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置，Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。   |
 | device   | -   | 指定 Tensor 的设备，Paddle 无此参数，一般对网络训练结果影响不大，可直接删除。  |
 | dtype   | -   | Tensor 的所需数据类型，Paddle 无此参数，一般对网络训练结果影响不大，可直接删除。 |
 | -             |weight_ih_attr| weight_ih 的参数， PyTorch 无此参数， Paddle 保持默认即可。  |
@@ -28,19 +28,17 @@ paddle.nn.GRUCell(input_size, hidden_size, weight_ih_attr=None, weight_hh_attr=N
 #### bias：是否使用偏置
 ```python
 # PyTorch 写法
-torch.nn.GRU(16, 32, bias=True)
-
+torch.nn.GRUCell(16, 32, bias=True)
 # Paddle 写法
-paddle.nn.GRU(16, 32)
+paddle.nn.GRUCell(16, 32)
 ```
 ```python
 # PyTorch 写法
-torch.nn.GRU(16, 32, bias=False)
-
+torch.nn.GRUCell(16, 32, bias=False)
 # Paddle 写法
-paddle.nn.GRU(16, 32, bias_ih_attr=False, bias_hh_attr=False)
+paddle.nn.GRUCell(16, 32, bias_ih_attr=False, bias_hh_attr=False)
 ```
-#### forward：前向传播
+#### forward 类方法：前向传播
 ```python
 # PyTorch 写法
 rnn = torch.nn.GRUCell(2, 2)
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTM.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTM.md
index 014f18a2d19..4584fdf6908 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTM.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTM.md
@@ -39,7 +39,7 @@ paddle.nn.LSTM(input_size,
 | input_size          | input_size            | 表示输入 x 的大小。  |
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
 | num_layers          | num_layers            | 表示循环网络的层数。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。   |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。   |
 | batch_first   | time_major   | PyTorch 表示 batch size 是否为第一维，PaddlePaddle 表示 time steps 是否为第一维，它们的意义相反。需要转写。 |
 | dropout   | dropout   | 表示 dropout 概率。  |
 | bidirectional | direction    | PyTorch 表示是否进行双向，Paddle 使用字符串表示是双向 LSTM（`bidirectional`）还是单向 LSTM（`forward`）|
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md
index f3698d4ccb3..54adb123931 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.LSTMCell.md
@@ -1,4 +1,4 @@
-## [ 输入参数用法不一致 ]torch.nn.LSTMCell
+## [ 返回参数类型不一致 ]torch.nn.LSTMCell
 ### [torch.nn.LSTMCell](https://pytorch.org/docs/stable/generated/torch.nn.LSTMCell.html#torch.nn.LSTMCell)
 ```python
 torch.nn.LSTMCell(input_size, hidden_size, bias=True, device=None, dtype=None)
@@ -9,14 +9,14 @@ torch.nn.LSTMCell(input_size, hidden_size, bias=True, device=None, dtype=None)
 paddle.nn.LSTMCell(input_size, hidden_size, weight_ih_attr=None, weight_hh_attr=None, bias_ih_attr=None, bias_hh_attr=None, proj_size=0, name=None)
 ```
 
-两者功能一致但参数不一，且二者的 forward 函数返回值格式不一致, 具体如下：
+两者功能一致但输入参数用法不一致，且返回参数类型不同，具体如下：
 ### 参数映射
 
 | PyTorch       | PaddlePaddle | 备注                                                   |
 | ------------- | ------------ | ------------------------------------------------------ |
 | input_size          | input_size            | 表示输入 x 的大小。  |
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。   |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。   |
 | device   | -   | 指定 Tensor 的设备，Paddle 无此参数，一般对网络训练结果影响不大，可直接删除。  |
 | dtype   | -   | Tensor 的所需数据类型，一般对网络训练结果影响不大，可直接删除。  |
 | -             |weight_ih_attr| weight_ih 的参数， PyTorch 无此参数， Paddle 保持默认即可。  |
@@ -29,19 +29,16 @@ paddle.nn.LSTMCell(input_size, hidden_size, weight_ih_attr=None, weight_hh_attr=
 ```python
 # PyTorch 写法
 torch.nn.LSTMCell(16, 32, bias=True)
-
 # Paddle 写法
 paddle.nn.LSTMCell(16, 32)
 ```
 ```python
 # PyTorch 写法
 torch.nn.LSTMCell(16, 32, bias=False)
-
 # Paddle 写法
 paddle.nn.LSTMCell(16, 32, bias_ih_attr=False, bias_hh_attr=False)
 ```
-
-#### forward：前向传播
+#### forward 类方法：前向传播
 ```python
 # PyTorch 写法
 rnn = torch.nn.LSTMCell(2, 2)
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNN.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNN.md
index aab6b8ea854..54095b40d53 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNN.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNN.md
@@ -27,7 +27,7 @@ paddle.nn.SimpleRNN(input_size, hidden_size, num_layers=1, activation='tanh', di
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
 | num_layers          | num_layers            | 表示循环网络的层数。  |
 | nonlinearity          | activation            | 表示激活函数类型，仅参数名不一致。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。   |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。   |
 | batch_first   | time_major   | PyTorch 表示 batch size 是否为第一维，PaddlePaddle 表示 time steps 是否为第一维，它们的意义相反。需要转写。  |
 | dropout   | dropout   | 表示 dropout 概率。  |
 | bidirectional | direction    | PyTorch 表示是否进行双向 RNN，Paddle 使用字符串表示是双向 RNN（`bidirectional`）还是单向 RNN（`forward`）。 |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNBase.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNBase.md
index 7b59964bec0..d4e45375bdf 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNBase.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNBase.md
@@ -24,7 +24,7 @@ paddle.nn.layer.rnn.RNNBase(mode, input_size, hidden_size,
 | input_size          | input_size            | 表示输入 x 的大小。  |
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
 | num_layers          | num_layers            | 表示循环网络的层数。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。   |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。   |
 | batch_first   | time_major   | PyTorch 表示 batch size 是否为第一维，PaddlePaddle 表示 time steps 是否为第一维，它们的意义相反。需要转写。  |
 | dropout   | dropout   | 表示 dropout 概率。  |
 | bidirectional | direction    | PyTorch 表示是否进行双向 RNN，Paddle 使用字符串表示是双向 RNN（`bidirectional`）还是单向 RNN（`forward`）。 |
diff --git a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md
index 90a887676b1..7aa872d7ff6 100644
--- a/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md
+++ b/docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.RNNCell.md
@@ -1,4 +1,4 @@
-## [ 输入参数用法不一致 ]torch.nn.RNNCell
+## [ 返回参数类型不一致 ]torch.nn.RNNCell
 ### [torch.nn.RNNCell](https://pytorch.org/docs/stable/generated/torch.nn.RNNCell.html#torch.nn.RNNCell)
 ```python
 torch.nn.RNNCell(input_size, hidden_size, bias=True, nonlinearity='tanh', device=None, dtype=None)
@@ -9,14 +9,14 @@ torch.nn.RNNCell(input_size, hidden_size, bias=True, nonlinearity='tanh', device
 paddle.nn.SimpleRNNCell(input_size, hidden_size, activation='tanh', weight_ih_attr=None, weight_hh_attr=None, bias_ih_attr=None, bias_hh_attr=None, name=None)
 ```
 
-两者功能一致但参数不一致，部分参数名不同，且二者的 forward 函数返回值格式不一致, 具体如下：
+两者功能一致但输入参数用法不一致，且返回参数类型不同，具体如下：
 ### 参数映射
 
 | PyTorch       | PaddlePaddle | 备注                                                   |
 | ------------- | ------------ | ------------------------------------------------------ |
 | input_size          | input_size            | 表示输入 x 的大小。  |
 | hidden_size          | hidden_size            | 表示隐藏状态 h 大小。  |
-| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 支持自定义偏置属性， torch 不支持，需要转写。   |
+| bias          | bias_ih_attr, bias_hh_attr  | 是否使用偏置， Paddle 的 bias_ih_attr, bias_hh_attr 参数均需与 PyTorch 设置一致，需要转写。   |
 | nonlinearity          | activation            | 表示激活函数类型，仅参数名不一致。  |
 | -             |weight_ih_attr| weight_ih 的参数， PyTorch 无此参数， Paddle 保持默认即可。  |
 | -             |weight_hh_attr| weight_hh 的参数，  PyTorch 无此参数， Paddle 保持默认即可。  |
@@ -26,19 +26,16 @@ paddle.nn.SimpleRNNCell(input_size, hidden_size, activation='tanh', weight_ih_at
 ```python
 # PyTorch 写法
 torch.nn.RNNCell(16, 32, bias=True)
-
 # Paddle 写法
 paddle.nn.SimpleRNNCell(16, 32)
 ```
 ```python
 # PyTorch 写法
 torch.nn.RNNCell(16, 32, bias=False)
-
 # Paddle 写法
 paddle.nn.SimpleRNNCell(16, 32, bias_ih_attr=False, bias_hh_attr=False)
 ```
-
-#### forward：前向传播
+#### forward 类方法：前向传播
 ```python
 # PyTorch 写法
 rnn = torch.nn.RNNCell(2, 2)