【Paddle Tensor 第二期 API支持 0-size Tensor】paddle.full_like 支持 0-size tensor-part #70077
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
你的PR提交成功，感谢你对开源项目的贡献! |
Comment on lines
42
to
44
| if (out->numel() == 0) { | ||
| return; // Handle 0-size Tensor: no allocation or computation needed. | ||
| } |
There was a problem hiding this comment.
这里得加一行红框里的代码：
其他地方也同理，不然没有place信息了
Comment on lines
+106
to
+108
| if (out->numel() == 0) { | ||
| return; | ||
| } |
Comment on lines
55
to
58
| out->Resize(x.dims()); | ||
| if (out->numel() == 0) { | ||
| return; | ||
| } |
There was a problem hiding this comment.
- out->Resize是新加的逻辑，放到if分支里
- if分支里缺少Alloc
| @@ -105,12 +105,12 @@ void FullBatchSizeLikeKernel(const Context& dev_ctx, | |||
| int out_batch_size_dim, | |||
| DenseTensor* out) { | |||
| if (x.lod().size() && x_batch_size_dim == 0) { | |||
| // set the correct batch size for the DenseTensor. | |||
Comment on lines
111
to
114
| } else { | ||
| FullLikeKernel<T, Context>(dev_ctx, x, val, dtype, out); | ||
| } | ||
| FullLikeKernel<T, Context>(dev_ctx, x, val, dtype, out); | ||
| } |
Comment on lines
224
to
225
| shape = [2, 3] | ||
| value = 5 |
There was a problem hiding this comment.
这里shape不是一个0-size，这个单测的意义是？
| ) | ||
| def test_full_kernel_gpu(self): | ||
| paddle.disable_static() | ||
| shape = [2, 3] |
There was a problem hiding this comment.
这个单测文件有测试0-size的case吗？
HydrogenSulfate
changed the title
@SCUcookie 这部分单侧写法下一个PR修改
|
HydrogenSulfate
approved these changes
Dec 13, 2024
Comment on lines
+221
to
+269
| class TestFullKernelZeroSize(unittest.TestCase): | ||
| def test_full_kernel_cpu_zero_size(self): | ||
| paddle.disable_static() | ||
| value = 5 | ||
| dtype = "int32" | ||
| shape = [0, 3] | ||
| tensor = paddle.full(shape, value, dtype=dtype) | ||
| expected = np.full(shape, value, dtype=dtype) | ||
| self.assertTrue(np.array_equal(tensor.numpy(), expected)) | ||
| paddle.enable_static() | ||
|
|
||
| @unittest.skipIf( | ||
| not core.is_compiled_with_cuda(), "Paddle is not compiled with CUDA" | ||
| ) | ||
| def test_full_kernel_gpu_zero_size(self): | ||
| paddle.disable_static() | ||
| paddle.set_device("gpu:0") | ||
| value = 5.5 | ||
| dtype = "float32" | ||
| shape = [0, 3] | ||
| tensor = paddle.full(shape, value, dtype=dtype) | ||
| expected = np.full(shape, value, dtype=dtype) | ||
| self.assertTrue(np.array_equal(tensor.numpy(), expected)) | ||
| paddle.enable_static() | ||
|
|
||
|
|
||
| class TestFullLikeKernelZeroSize(unittest.TestCase): | ||
| def test_full_like_kernel_cpu_zero_size(self): | ||
| paddle.disable_static() | ||
| base_tensor = paddle.to_tensor(np.empty((0, 2), dtype=np.float32)) | ||
| value = 10.0 | ||
| result = paddle.full_like(base_tensor, value, dtype="float32") | ||
| expected = np.full_like(base_tensor.numpy(), value) | ||
| self.assertTrue(np.array_equal(result.numpy(), expected)) | ||
| paddle.enable_static() | ||
|
|
||
| @unittest.skipIf( | ||
| not core.is_compiled_with_cuda(), "Paddle is not compiled with CUDA" | ||
| ) | ||
| def test_full_like_kernel_gpu_zero_size(self): | ||
| paddle.disable_static() | ||
| base_tensor = paddle.to_tensor( | ||
| np.empty((0, 3), dtype=np.float32), place=paddle.CUDAPlace(0) | ||
| ) | ||
| value = 20.0 | ||
| result = paddle.full_like(base_tensor, value, dtype="float32") | ||
| expected = np.full_like(base_tensor.numpy(), value) | ||
| self.assertTrue(np.array_equal(result.numpy(), expected)) | ||
| paddle.enable_static() |
There was a problem hiding this comment.
- enable/disable的切换使用guard
- self.assertTrue(np.array_equal/allclose(result.numpy(), expected))这类判断使用 np.testing.assert_array_equal/allclose
|
@SCUcookie 下个PR按照review意见，改一下单测的代码 |
74 tasks
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR Category
User Experience
PR Types
Bug fixes
Description
添加输入等于0时的情况判断