Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[Rocm] fix test of random_crop_op & logsumexp #32824

Merged
merged 7 commits into from
May 19, 2021
Merged

[Rocm] fix test of random_crop_op & logsumexp #32824

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
3055394
[Rocm] fix test of random_crop_op
Ray2020BD May 10, 2021
9eec9e6
[Rocm] fix test of random_crop_op
Ray2020BD May 13, 2021
62556c4
[Rocm] fix test of random_crop_op & simple_rnn_op
Ray2020BD May 14, 2021
10e1cc2
[Rocm] fix test of random_crop_op & simple_rnn_op & logsumexp
Ray2020BD May 17, 2021
da04a55
[Rocm] fix test of random_crop_op & simple_rnn_op & logsumexp
Ray2020BD May 17, 2021
1dab597
[Rocm] fix test of random_crop_op & simple_rnn_op & logsumexp
Ray2020BD May 18, 2021
819915a
[Rocm] fix test of random_crop_op & logsumexp
Ray2020BD May 18, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions paddle/fluid/operators/random_crop_op.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,6 +59,7 @@ HOSTDEVICE inline void StridedMemcpy(const T* x, const size_t* x_dims, T* out,
size_t offset_i = offsets[i];

if (i == rank - 1) {
#ifndef PADDLE_WITH_HIP
PADDLE_ENFORCE(x_stride == 1,
"When i:%d == rank:%d - 1, x_stride of random_crop_op "
"expected to be 1, but got %ld. Please check input "
Expand All @@ -69,6 +70,7 @@ HOSTDEVICE inline void StridedMemcpy(const T* x, const size_t* x_dims, T* out,
"expected to be 1, but got %ld. Please check input "
"value.",
i, rank, out_stride);
#endif
Copy link
Contributor

@qili93 qili93 May 13, 2021 •
edited
Loading

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

看了下你的分析文档，但当前这个修复方式不能满足OP要，我们OP修复的原则是

  • 不可以因为这两段代码会挂就不检查x_stride和out_stride，而是需要找到能工作的代码替换这两段代码，或者修复PADDLE_ENFORCE的逻辑使其能够工作

我们修复OP的目的不是为了单测可以通过，而是保证功能的正确性，当前的diff只是简单跳过了x_stride和out_stride的检查，不符合我们OP修复的目的。

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

好的，已修改。

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这个直接printf不符合paddle msg的标准，请参考这个文档写关于paddle的error msg
https://github.com/PaddlePaddle/Paddle/wiki/Paddle-Error-Message-Writing-Specification

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

已删除

x += offset_i;
for (size_t j = 0; j < out_dim_i; ++j) {
*out++ = *x++;
Expand Down
13 changes: 11 additions & 2 deletions python/paddle/fluid/tests/unittests/test_simple_rnn_op.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,8 +44,11 @@ def get_weight_names(self):

def setUp(self):
self.op_type = "rnn"
self.dtype = np.float64
self.sequence_length = np.array([12, 11, 10, 9, 8], dtype=np.int32)
self.dtype = np.float32 if fluid.core.is_compiled_with_rocm(
) else np.float64
self.sequence_length = None if fluid.core.is_compiled_with_rocm(
) else np.array(
[12, 11, 10, 9, 8], dtype=np.int32)
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

MIOPEN不支持sequence_length的话，SimpleRNN的API也同样需要修改，对于sequence_length非空的输入需要抛出正确的error msg.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

simpleRNN的修改以及该单测的修改从本pr中删除，将和其他rnn单测修改统一放到其他的pr中。

self.num_layers = 1
self.is_bidirec = False
self.is_test = False
Expand Down Expand Up @@ -82,6 +85,12 @@ def setUp(self):

output, last_hidden = rnn1(input, sequence_length=self.sequence_length)

if fluid.core.is_compiled_with_rocm():
self._get_places = lambda: [fluid.core.CUDAPlace(0)]
if self.is_bidirec:
for i in range(0, len(flat_w), 4):
flat_w[i + 1], flat_w[i + 2] = flat_w[i + 2], flat_w[i + 1]

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

这里能解释一下为什么需要对flat_w的数据做交换吗？是miopenSetRNNDescriptor_V2 里面的 miopenRNNbidirection 和 cudnnSetRNNDescriptor_v6 的 CUDNN_BIDIRECTIONAL 的行为不一致？

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

miopen相关api的实现中，layout并没有和cuda相关pai对齐，需要手动调节。

init_h = np.zeros((self.num_layers * self.direction_num, batch_size,
hidden_size)).astype(self.dtype)

Expand Down