映射文档 torch.nn.functional.embedding by LokeZhou · Pull Request #5928 · PaddlePaddle/docs

LokeZhou · 2023-06-09T08:12:20Z

add document function.embedding

paddle-bot · 2023-06-09T08:12:24Z

感谢你贡献飞桨文档，文档预览构建中，Docs-New 跑完后即可预览，预览链接：http://preview-pr-5928.paddle-docs-preview.paddlepaddle.org.cn/documentation/docs/zh/api/index_cn.html
预览工具的更多说明，请参考：飞桨文档预览工具

zhwesky2010 · 2023-06-13T09:40:58Z

...t/convert_from_pytorch/api_difference/functional/torch.nn.functional.binary_cross_entropy.md

@@ -0,0 +1,42 @@
+## torch.nn.functional.binary_cross_entropy


前面要写分类

zhwesky2010 · 2023-06-13T09:41:39Z

...odel_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.embedding.md

+| padding_idx     | padding_idx            | 在此区间内的参数及对应的梯度将会以 0 进行填充  |
+| max_norm      | -            | 如果给定，Embeddding 向量的范数（范数的计算方式由 norm_type 决定）超过了 max_norm 这个界限，就要再进行归一化，PaddlePaddle 无此功能，暂无转写方式。  |
+| norm_type     | -            | 为 maxnorm 选项计算 p-范数的 p。默认值 2，PaddlePaddle 暂无此功能，暂无转写方式。  |
+| scale_grad_by_freq | -       | 是否根据单词在 mini-batch 中出现的频率，对梯度进行放缩，PaddlePaddle 暂无此功能。  |


少了个暂无转写方式。

写成统一一致的吧：暂无转写方式

zhwesky2010 · 2023-06-13T09:41:51Z

...odel_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.embedding.md

+| norm_type     | -            | 为 maxnorm 选项计算 p-范数的 p。默认值 2，PaddlePaddle 暂无此功能，暂无转写方式。  |
+| scale_grad_by_freq | -       | 是否根据单词在 mini-batch 中出现的频率，对梯度进行放缩，PaddlePaddle 暂无此功能。  |
+| sparse     | sparse            | 表示是否使用稀疏更新。  |
+| -             | name  | Pytorch 无此参数，Paddle 保持默认即可。  |


所有API的name参数，都忽略即可

zhwesky2010 · 2023-06-13T09:42:30Z

CI要检查一下

zhwesky2010 · 2023-06-20T09:52:38Z

...t/convert_from_pytorch/api_difference/functional/torch.nn.functional.binary_cross_entropy.md

@@ -0,0 +1,42 @@
+## [ 参数不一致 ] torch.nn.functional.binary_cross_entropy


这个是torch参数更多吧

zhwesky2010 · 2023-06-20T09:53:12Z

...odel_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.embedding.md

+| padding_idx     | padding_idx            | 在此区间内的参数及对应的梯度将会以 0 进行填充  |
+| max_norm      | -            | 如果给定，Embeddding 向量的范数（范数的计算方式由 norm_type 决定）超过了 max_norm 这个界限，就要再进行归一化，PaddlePaddle 无此功能，暂无转写方式。  |
+| norm_type     | -            | 为 maxnorm 选项计算 p-范数的 p。默认值 2，PaddlePaddle 暂无此功能，暂无转写方式。  |
+| scale_grad_by_freq | -       | 是否根据单词在 mini-batch 中出现的频率，对梯度进行放缩，PaddlePaddle 暂无此功能。  |


写成统一一致的吧：暂无转写方式

zhwesky2010 · 2023-06-20T09:53:59Z

docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.Embedding.md

@@ -33,3 +33,21 @@ paddle.nn.Embedding(num_embeddings,
 | scale_grad_by_freq | -       | 是否根据单词在 mini-batch 中出现的频率，对梯度进行放缩，PaddlePaddle 暂无此功能。  |


这些都用统一的话术描述：Paddle无此功能，暂无转写方式。

zhwesky2010 · 2023-06-20T09:56:38Z

...s/model_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.unfold.md

+| ------------- | ------------ | ------------------------------------------------------ |
+| input     | x            | 输入 Tensor  |
+| kernel_size   | kernel_sizes | 卷积核大小， PyTorch 参数类型为 int、tuple(int) 或者 list(int)， Paddle 参数类型为 int 或者 list(int)。   |
+| dilation      | dilations    | 卷积膨胀，PyTorch 参数类型为 int、tuple(int) 或者 list(int)， Paddle 参数类型为 int 或者 list(int)。 |


可以写成，Pytorch支持int、tuple(int) 或者 list(int)，Paddle仅支持int 或者 list(int)，突出我们比他们少

zhwesky2010

需要写是在转写哪个参数，参考模板来

zhwesky2010 · 2023-06-26T14:03:24Z

...odel_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.embedding.md

+| sparse     | sparse            | 表示是否使用稀疏更新。  |
+
+### 转写示例
+```python


转写前需要写下是转写哪个参数，请严格按规范来

那这个是转哪个参数呢

无参数转写，已删除

zhwesky2010 · 2023-06-26T14:03:30Z

...s/model_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.unfold.md

-| padding     | paddings     | 每个维度的扩展，仅参数名不一致。  |
-| stride      | strides      | 卷积步长，仅参数名不一致。        |
+### 转写示例
+``` python


转写前需要写下是转写哪个参数

zhwesky2010 · 2023-06-26T14:03:36Z

docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.Embedding.md

+
+
+### 转写示例
+```python


转写前需要写下是转写哪个参数，请严格按规范来

zhwesky2010 · 2023-07-04T09:45:37Z

...odel_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.embedding.md

+| sparse     | sparse            | 表示是否使用稀疏更新。  |
+
+### 转写示例
+```python


那这个是转哪个参数呢

zhwesky2010 · 2023-07-04T09:46:35Z

...s/model_convert/convert_from_pytorch/api_difference/functional/torch.nn.functional.unfold.md

-## [参数名不一致]torch.nn.functional.unfold

-### [torch.nn.functional.unfold](https://pytorch.org/docs/1.13/generated/torch.nn.functional.unfold.html#torch.nn.functional.unfold)
+## [ 参数名不一致 ]torch.nn.functional.unfold


这个就算 参数不一致 吧，支持类型不同

zhwesky2010 · 2023-07-04T09:47:31Z

docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.Embedding.md

@@ -28,8 +28,27 @@ paddle.nn.Embedding(num_embeddings,
 | num_embeddings     | num_embeddings            | 表示嵌入字典的大小。  |
 | embedding_dim     | embedding_dim            | 表示每个嵌入向量的维度。  |
 | padding_idx     | padding_idx            | 在此区间内的参数及对应的梯度将会以 0 进行填充  |


对于需要转写的参数，需要加一个需要转写

zhwesky2010 · 2023-07-04T09:48:07Z

docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.Embedding.md

+# paddle 写法
+paddle.nn.Embedding(4, 3,padding_idx=0)
+
+# 当 PyTorch 使用 max_norm、norm_type、scale_grad_by_freq 任意参数时，暂不支持转写


这个不用写

zhwesky2010 · 2023-07-04T09:48:52Z

docs/guides/model_convert/convert_from_pytorch/api_difference/nn/torch.nn.Embedding.md

+torch.nn.Embedding(4, 3)
+
+# paddle 写法
+paddle.nn.Embedding(4，3)


这个不需要转写的

转写示例是针对某个参数，抬头说明是要转哪个参数

不需要转写的情况不要写

add convert_from_pytoch function.embedding md

dd20f51

LokeZhou mentioned this pull request Jun 9, 2023

Add embedding PaddlePaddle/PaConvert#94

Merged

add binary_cross_entropy and unfold doc

024994a

zhwesky2010 reviewed Jun 13, 2023

View reviewed changes

LokeZhou added 2 commits June 16, 2023 02:09

add exp

2c3077a

fix function unfold

b1ed752

LokeZhou mentioned this pull request Jun 19, 2023

add some loss api PaddlePaddle/PaConvert#109

Merged

zhwesky2010 reviewed Jun 20, 2023

View reviewed changes

zhwesky2010 changed the title ~~add document function.embedding~~ 映射关系 add document function.embedding Jun 20, 2023

zhwesky2010 changed the title ~~映射关系 add document function.embedding~~ 映射文档 add document function.embedding Jun 20, 2023

zhwesky2010 changed the title ~~映射文档 add document function.embedding~~ 映射文档 torch.nn.function.embedding Jun 20, 2023

zhwesky2010 changed the title ~~映射文档 torch.nn.function.embedding~~ 映射文档 torch.nn.functional.embedding Jun 20, 2023

LokeZhou added 4 commits June 20, 2023 11:48

fix comment

1cec3cd

Merge branch 'develop' into add_embedding

e6d3e11

fix unfold ci

d18c28f

fix unfold ci

48cbfa6

zhwesky2010 reviewed Jun 26, 2023

View reviewed changes

add args

8d51079

zhwesky2010 reviewed Jul 4, 2023

View reviewed changes

LokeZhou added 2 commits July 6, 2023 09:38

delete some exp

5d047de

Merge branch 'develop' into add_embedding

d1c42c0

		@@ -0,0 +1,42 @@
		## [ 参数不一致 ] torch.nn.functional.binary_cross_entropy

		@@ -33,3 +33,21 @@ paddle.nn.Embedding(num_embeddings,
		\| scale_grad_by_freq \| - \| 是否根据单词在 mini-batch 中出现的频率，对梯度进行放缩，PaddlePaddle 暂无此功能。 \|



		### 转写示例
		```python

Conversation

LokeZhou commented Jun 9, 2023

Uh oh!

paddle-bot bot commented Jun 9, 2023

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

zhwesky2010 commented Jun 13, 2023

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

zhwesky2010 left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

zhwesky2010 Jul 4, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

zhwesky2010 left a comment •

edited

Loading

zhwesky2010 Jul 4, 2023 •

edited

Loading