mirror of
https://github.com/zebrajr/pytorch.git
synced 2025-12-07 12:21:27 +01:00
6706e9af19
Summary: Pull Request resolved: https://github.com/pytorch/pytorch/pull/17481 Usually, feature macros are either defined or undefined and checked accordingly. C10_MOBILE was a weird special case that was always defined but either defined to 1 or to 0. This caused a lot of confusion for me when trying to disable something from mobile build and it also disabled it from the server build (because I was using ifdef). Also, I found a place in the existing code base that made that wrong assumption and used the macro wrongly, see https://fburl.com/y4icohts Reviewed By: dzhulgakov Differential Revision: D14214825 fbshipit-source-id: f3a155b6d43d334e8839e2b2e3c40ed2c773eab6
46 lines
1.4 KiB
C++
46 lines
1.4 KiB
C++
#ifndef CAFFE2_OPERATORS_CONV_TRANSPOSE_MOBILE_OP_H_
|
|
#define CAFFE2_OPERATORS_CONV_TRANSPOSE_MOBILE_OP_H_
|
|
|
|
#include "caffe2/core/common.h"
|
|
|
|
#ifdef C10_MOBILE
|
|
|
|
#include "caffe2/core/context.h"
|
|
#include "caffe2/core/operator.h"
|
|
#include "caffe2/operators/conv_transpose_unpool_op_base.h"
|
|
|
|
namespace caffe2 {
|
|
|
|
template <typename T, class Context>
|
|
class ConvTransposeMobileOp final : public ConvTransposeUnpoolBase<Context> {
|
|
public:
|
|
USE_CONV_TRANSPOSE_UNPOOL_BASE_FUNCTIONS(Context);
|
|
ConvTransposeMobileOp(const OperatorDef& operator_def, Workspace* ws)
|
|
: ConvTransposeUnpoolBase<Context>(operator_def, ws) {
|
|
OPERATOR_NEEDS_FEATURE(order_ == StorageOrder::NCHW, "Only NCHW order is supported right now.");
|
|
OPERATOR_NEEDS_FEATURE(
|
|
this->pad_l() == 0, "operator does not handle row width padding");
|
|
OPERATOR_NEEDS_FEATURE(
|
|
this->pad_r() == 0, "operator does not handle row width padding");
|
|
OPERATOR_NEEDS_FEATURE(this->stride_w() <= 4, "stride width must be <= 4");
|
|
}
|
|
|
|
bool RunOnDeviceWithOrderNCHW() override;
|
|
bool RunOnDeviceWithOrderNHWC() override;
|
|
|
|
private:
|
|
// We store a numThreasds per-worker tiles of Y, and numThreads per-worker threadBuffer for the
|
|
// gemm output, laid out in that order.
|
|
Tensor threadBuffer_{CPU};
|
|
|
|
// Input: X, W, b
|
|
// Output: Y
|
|
INPUT_TAGS(INPUT, FILTER, BIAS);
|
|
};
|
|
|
|
} // namespace caffe2
|
|
|
|
#endif // C10_MOBILE
|
|
|
|
#endif // CAFFE2_OPERATORS_CONV_TRANSPOSE_MOBILE_OP_H_
|