【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part #6212

Li-fAngyU · 2023-09-26T03:30:13Z

【PaddlePaddle Hackathon 5th】开源贡献个人挑战赛 Paddle#57262

注: 对于torch.Tensor.resize_(*sizes, memory_format=torch.contiguous_format) 这个API，因为torch.Tensor.resize_支持输入size比原先Tensor更大或更小的维度，而目前Paddle没有该对应功能的实现。

# Torch code:
x = torch.tensor([[1, 2], [3, 4], [5, 6]])
x.resize_(2, 2)
# tensor([[ 1,  2],
#         [ 3,  4]])

paddle-bot · 2023-09-26T03:30:18Z

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-6212.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

paddle-bot · 2023-09-26T03:30:22Z

✅ This PR's description meets the template requirements!
Please wait for other CI results.

docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.as_strided.md

docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.index_put.md

docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.view_as.md

docs/guides/model_convert/convert_from_pytorch/api_difference/ops/torch.as_strided.md

zhwesky2010 · 2023-10-11T10:43:15Z

docs/guides/model_convert/convert_from_pytorch/pytorch_api_mapping_cn.md

@@ -917,7 +917,9 @@
 | 300 | [torch.ShortTensor](https://pytorch.org/docs/stable/tensors.html)                                                                                                                           |[paddle.to_tensor](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/to_tensor_cn.html#to-tensor)| 仅 paddle 参数更多 , [差异对比](https://github.com/PaddlePaddle/docs/tree/develop/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.ShortTensor.md)|
 | 301 |  [torch.Tensor.copysign](https://pytorch.org/docs/1.13/generated/torch.Tensor.copysign.html?highlight=torch+tensor+copysign#torch.Tensor.copysign)  |  | 功能缺失        |
 | 302 |  [torch.Tensor.dequantize](https://pytorch.org/docs/1.13/generated/torch.Tensor.dequantize.html?highlight=torch+tensor+dequantize#torch.Tensor.dequantize)  |  | 功能缺失        |
-
+| 303 |  [torch.Tensor.is_pinned](https://pytorch.org/docs/stable/generated/torch.Tensor.is_pinned.html?highlight=is_pinned#torch.Tensor.is_pinned)  |  |    功能缺失  |


这个有办法组合替代实现吧，比如说判断x.place

谢谢审查，经过查阅资料是可以组合替代实现判断张量是否在固定内存上，但是Paddle和Torch的pin_memory有一点不太一致。

Paddle 代码：

pin_memory_Tensor = paddle.to_tensor(1, place=paddle.CUDAPinnedPlace()) print(pin_memory_Tensor.place) # Place(gpu_pinned)

Torch 代码：

x = torch.randn(4,4) print(x.pin_memory().is_pinned()) # True print(x.cuda().pin_memory().is_pinned()) # 会报错 # RuntimeError: cannot pin 'torch.cuda.FloatTensor' only dense CPU tensors can be pinned

即 Paddle 仅支持 CUDA Pin，而 Torch 仅支持 dense CPU tensors to pin memory。

请问是否忽略这一点呢？

这个点不用管，只需关注torch的代码转换为paddle如何呈现（前提是torch代码可正常运行），不关注其他的地方。上文中的 x.is_pinned() 如果转为 paddle的写法，'pinned' in x.place 这种写法有问题吗

好的，可以通过 'pinned' in str(x.place) 去判断。

zhwesky2010 · 2023-10-11T10:43:47Z

docs/guides/model_convert/convert_from_pytorch/pytorch_api_mapping_cn.md

@@ -917,7 +917,9 @@
 | 300 | [torch.ShortTensor](https://pytorch.org/docs/stable/tensors.html)                                                                                                                           |[paddle.to_tensor](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/to_tensor_cn.html#to-tensor)| 仅 paddle 参数更多 , [差异对比](https://github.com/PaddlePaddle/docs/tree/develop/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.ShortTensor.md)|
 | 301 |  [torch.Tensor.copysign](https://pytorch.org/docs/1.13/generated/torch.Tensor.copysign.html?highlight=torch+tensor+copysign#torch.Tensor.copysign)  |  | 功能缺失        |
 | 302 |  [torch.Tensor.dequantize](https://pytorch.org/docs/1.13/generated/torch.Tensor.dequantize.html?highlight=torch+tensor+dequantize#torch.Tensor.dequantize)  |  | 功能缺失        |
-
+| 303 |  [torch.Tensor.is_pinned](https://pytorch.org/docs/stable/generated/torch.Tensor.is_pinned.html?highlight=is_pinned#torch.Tensor.is_pinned)  |  |    功能缺失  |
+| 304 |  [torch.Tensor.sum_to_size](https://pytorch.org/docs/stable/generated/torch.Tensor.sum_to_size.html?highlight=sum_to_size#torch.Tensor.sum_to_size)  |  |    功能缺失  |


这个没办法组合替代实现吗

torch.sum_to_size 的实现是借助 sum_to 实现的其源码如下:

template <typename T> inline Tensor _sum_to( Tensor tensor, const c10::ArrayRef<T> shape, bool always_return_non_view = false) { if (shape.size() == 0) { return tensor.sum(); } auto sizes = at::symint::sizes<T>(tensor); c10::SmallVector<int64_t, 8> reduce_dims; const int64_t leading_dims = sizes.size() - shape.size(); for (const auto i : c10::irange(leading_dims)) { reduce_dims.push_back(i); } for (int64_t i = leading_dims; i < static_cast<int64_t>(sizes.size()); ++i) { if (shape[i - leading_dims] == 1 && sizes[i] != 1) { reduce_dims.push_back(i); } } if (!reduce_dims.empty()) { tensor = tensor.sum(reduce_dims, /*keepdim=*/true); } if (always_return_non_view) { // This is only actually used by the functionalization pass. // We want to be able to guarantee that this function doesn't return a view // of the input. return leading_dims > 0 ? at::symint::view_copy<T>(tensor, shape) : tensor.clone(); } else { return leading_dims > 0 ? at::symint::view<T>(tensor, shape) : tensor; } }

由于 Paddle 中没有 sum_to API 因此我不知道该如何去组合替代实现 sum_to_size.

这个就先按功能缺失来吧

zhwesky2010

上面有些评论看起来还没有改，但是mark_resolved了，有些描述不符合模板规范

Li-fAngyU · 2023-10-17T10:34:37Z

抱歉疏忽了，已修改！

zhwesky2010 · 2023-10-17T13:29:06Z

docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.Tensor.view.md

+
+| PyTorch | PaddlePaddle | 备注                           |
+| ------- | ------------ | ------------------------------ |
+| *shape   | shape_or_dtype         | 指定的维度。 Pytorch 参数 shape 既可以是可变参数，也可以是 list/tuple/torch.Size/dtype 的形式， Paddle 参数 shape_or_dtype 为 list/tuple/dtype 的形式。需要进行转写。 |


对于可变参数的用法，需要进行转写

zhwesky2010 · 2023-10-17T13:30:20Z

docs/guides/model_convert/convert_from_pytorch/api_difference/ops/torch.as_strided.md

+### 参数映射
+| PyTorch       | PaddlePaddle | 备注                                                   |
+| ------------- | ------------ | ------------------------------------------------------ |
+| input          | x         | 表示输入的 Tensor 。                                     |


你这个写了 仅参数名不一致 吗？

zhwesky2010 · 2023-10-17T13:31:32Z

docs/guides/model_convert/convert_from_pytorch/api_difference/ops/torch.cummax.md

+torch.cummax(x,1, out=(values, indices))
+
+# Paddle 写法
+paddle.assign(paddle.cummax(x,1), (values, indices))


这个需要用两行转写吧

这里是参考了其他文档, 比如 torch.histc.md。

# Pytorch 写法 torch.histc(x, out=y) # Paddle 写法 paddle.assign(paddle.histogram(x).astype('float32'), y)

zhwesky2010 · 2023-10-17T13:32:07Z

docs/guides/model_convert/convert_from_pytorch/api_difference/ops/torch.nanquantile.md

+|   q   |  q  | 待计算的分位数。  |
+|  dim  | axis| 指定对 x 进行计算的轴，仅参数名不一致。 |
+|keepdim|keepdim| 是否在输出 Tensor 中保留减小的维度。|
+|interpolation|  - | 当所需分位数位于两个数据点之间时使用的插值方法，Paddle 无此参数，需要转写，Paddle 暂无转写方式。|


这个 需要转写 不用写

zhwesky2010 · 2023-10-17T13:32:42Z

docs/guides/model_convert/convert_from_pytorch/api_difference/ops/torch.nanquantile.md

+#### out：指定输出
+```python
+# Pytorch 写法
+torch.nanquantile(torch.tensor([float('nan'), 1., 2., 3.]), 0.6, interpolation='linear', out=y)


interpolation 这个不能转写，写示例不用加上

zhwesky2010 · 2023-10-17T13:33:09Z

docs/guides/model_convert/convert_from_pytorch/pytorch_api_mapping_cn.md

@@ -917,7 +917,9 @@
 | 300 | [torch.ShortTensor](https://pytorch.org/docs/stable/tensors.html)                                                                                                                           |[paddle.to_tensor](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/to_tensor_cn.html#to-tensor)| 仅 paddle 参数更多 , [差异对比](https://github.com/PaddlePaddle/docs/tree/develop/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.ShortTensor.md)|
 | 301 |  [torch.Tensor.copysign](https://pytorch.org/docs/1.13/generated/torch.Tensor.copysign.html?highlight=torch+tensor+copysign#torch.Tensor.copysign)  |  | 功能缺失        |
 | 302 |  [torch.Tensor.dequantize](https://pytorch.org/docs/1.13/generated/torch.Tensor.dequantize.html?highlight=torch+tensor+dequantize#torch.Tensor.dequantize)  |  | 功能缺失        |
-
+| 303 |  [torch.Tensor.is_pinned](https://pytorch.org/docs/stable/generated/torch.Tensor.is_pinned.html?highlight=is_pinned#torch.Tensor.is_pinned)  |  |    功能缺失  |


这个已经给了你转写替代方案

zhwesky2010 · 2023-10-17T13:33:29Z

docs/guides/model_convert/convert_from_pytorch/pytorch_api_mapping_cn.md

@@ -917,7 +917,9 @@
 | 300 | [torch.ShortTensor](https://pytorch.org/docs/stable/tensors.html)                                                                                                                           |[paddle.to_tensor](https://www.paddlepaddle.org.cn/documentation/docs/zh/develop/api/paddle/to_tensor_cn.html#to-tensor)| 仅 paddle 参数更多 , [差异对比](https://github.com/PaddlePaddle/docs/tree/develop/docs/guides/model_convert/convert_from_pytorch/api_difference/Tensor/torch.ShortTensor.md)|
 | 301 |  [torch.Tensor.copysign](https://pytorch.org/docs/1.13/generated/torch.Tensor.copysign.html?highlight=torch+tensor+copysign#torch.Tensor.copysign)  |  | 功能缺失        |
 | 302 |  [torch.Tensor.dequantize](https://pytorch.org/docs/1.13/generated/torch.Tensor.dequantize.html?highlight=torch+tensor+dequantize#torch.Tensor.dequantize)  |  | 功能缺失        |
-
+| 303 |  [torch.Tensor.is_pinned](https://pytorch.org/docs/stable/generated/torch.Tensor.is_pinned.html?highlight=is_pinned#torch.Tensor.is_pinned)  |  |    功能缺失  |
+| 304 |  [torch.Tensor.sum_to_size](https://pytorch.org/docs/stable/generated/torch.Tensor.sum_to_size.html?highlight=sum_to_size#torch.Tensor.sum_to_size)  |  |    功能缺失  |


这个就先按功能缺失来吧

Li-fAngyU · 2023-10-18T02:42:38Z

Done! 已对上述问题进行了修改，其中针对 cummax 的转写示例还未改动。

zhwesky2010

LGTM

Li-fAngyU and others added 3 commits September 26, 2023 11:26

Hackthon5 No 42. paddle/docs first commit.

52028bc

first commit.

e46bf1f

Merge branch 'PaddlePaddle:develop' into develop

614b9ad

paddle-bot bot added the contributor label Sep 26, 2023

Li-fAngyU changed the title ~~【Hackthon 5】No.42 为Paddle代码转换工具新增API转换规则 (第1组编号1-20)~~ 【Hackthon 5 No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20) Sep 26, 2023

luotao1 assigned luotao1 and zhwesky2010 Sep 26, 2023

luotao1 added the PaddlePaddle Hackathon 飞桨黑客松活动issue与PR label Sep 26, 2023

Li-fAngyU added 3 commits September 26, 2023 18:12

remove empty file

1b437f1

Merge branch 'develop' of https://github.com/Li-fAngyU/docs into develop

ceb6d6e

fix torch.Tensor.cummin.md torch.Tensor.cummax.md.

f61679a

Li-fAngyU mentioned this pull request Sep 27, 2023

【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20) -part PaddlePaddle/PaConvert#306

Merged

zhwesky2010 reviewed Oct 11, 2023

View reviewed changes

Li-fAngyU and others added 2 commits October 11, 2023 20:51

Merge branch 'PaddlePaddle:develop' into develop

a25edd2

update

aab37c2

Li-fAngyU requested a review from zhwesky2010 October 11, 2023 13:47

zhwesky2010 reviewed Oct 17, 2023

View reviewed changes

fix torch.as_strided and torch.Tensor.as_strided

938fad1

zhwesky2010 reviewed Oct 17, 2023

View reviewed changes

update

a2c64ee

Li-fAngyU requested a review from zhwesky2010 October 18, 2023 03:01

zhwesky2010 approved these changes Oct 19, 2023

View reviewed changes

zhwesky2010 merged commit 84a39dd into PaddlePaddle:develop Oct 19, 2023
1 check passed

luotao1 changed the title ~~【Hackthon 5 No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)~~ 【Hackthon 5 No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part Oct 19, 2023

luotao1 changed the title ~~【Hackthon 5 No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part~~ 【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part Oct 19, 2023

Ligoml mentioned this pull request Oct 20, 2023

【PaddlePaddle Hackathon 5th】开源贡献个人挑战赛 PaddlePaddle/Paddle#57262

Open

Li-fAngyU mentioned this pull request Oct 23, 2023

【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20) PaddlePaddle/PaConvert#318

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part #6212

【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part #6212

Li-fAngyU commented Sep 26, 2023 •

edited

Loading

paddle-bot bot commented Sep 26, 2023

paddle-bot bot commented Sep 26, 2023 •

edited

Loading

zhwesky2010 Oct 11, 2023

Li-fAngyU Oct 11, 2023 •

edited

Loading

zhwesky2010 Oct 17, 2023

Li-fAngyU Oct 18, 2023

zhwesky2010 Oct 11, 2023

Li-fAngyU Oct 11, 2023

zhwesky2010 Oct 17, 2023

zhwesky2010 left a comment •

edited

Loading

Li-fAngyU commented Oct 17, 2023

zhwesky2010 Oct 17, 2023

zhwesky2010 Oct 17, 2023

zhwesky2010 Oct 17, 2023

Li-fAngyU Oct 18, 2023 •

edited

Loading

zhwesky2010 Oct 17, 2023

zhwesky2010 Oct 17, 2023

zhwesky2010 Oct 17, 2023

zhwesky2010 Oct 17, 2023

Li-fAngyU commented Oct 18, 2023

zhwesky2010 left a comment

【Hackathon 5th No.42】 为Paddle代码转换工具新增API转换规则 (第1组 编号1-20)-part #6212

【Hackathon 5th No.42】 为Paddle代码转换工具新增API转换规则 (第1组 编号1-20)-part #6212

Conversation

Li-fAngyU commented Sep 26, 2023 • edited Loading

paddle-bot bot commented Sep 26, 2023

paddle-bot bot commented Sep 26, 2023 • edited Loading

Choose a reason for hiding this comment

Li-fAngyU Oct 11, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

zhwesky2010 left a comment • edited Loading

Choose a reason for hiding this comment

Li-fAngyU commented Oct 17, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Li-fAngyU Oct 18, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Li-fAngyU commented Oct 18, 2023

zhwesky2010 left a comment

Choose a reason for hiding this comment

【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part #6212

【Hackathon 5th No.42】为Paddle代码转换工具新增API转换规则 (第1组编号1-20)-part #6212

Li-fAngyU commented Sep 26, 2023 •

edited

Loading

paddle-bot bot commented Sep 26, 2023 •

edited

Loading

Li-fAngyU Oct 11, 2023 •

edited

Loading

zhwesky2010 left a comment •

edited

Loading

Li-fAngyU Oct 18, 2023 •

edited

Loading