OSSForks/pytorch
1
0
Fork 0
mirror of https://github.com/zebrajr/pytorch.git synced 2025-12-06 12:20:52 +01:00
Code Issues Actions 38 Packages Projects Releases Wiki Activity

Revert "[ROCm] Implemented dropout usage for RNN with MIOpen backend (#144572)"

Browse Source 
This reverts commit e7bf490c43.

Reverted https://github.com/pytorch/pytorch/pull/144572 on behalf of https://github.com/jeanschmidt due to Broke internal signals, D69994027, I'll find someone to help get this change merged ([comment](https://github.com/pytorch/pytorch/pull/144572#issuecomment-2676314308))
This commit is contained in:
PyTorch MergeBot 2025-02-22 17:19:38 +00:00
parent 3409cbd177
commit bea72180ed
2 changed files with 5 additions and 87 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
aten/src/ATen/miopen/Descriptors.h
View File

@ -121,16 +121,6 @@ struct ConvolutionDescriptor
}
};
struct DropoutDescriptor
: public Descriptor<miopenDropoutDescriptor,
&miopenCreateDropoutDescriptor,
&miopenDestroyDropoutDescriptor>
{
void set(miopenHandle_t handle, float dropout, void* states, size_t stateSizeInBytes,
unsigned long long seed, bool use_mask, bool state_evo, miopenRNGType_t rng_mode) {
MIOPEN_CHECK(miopenSetDropoutDescriptor(mut_desc(), handle, dropout, states, stateSizeInBytes, seed, use_mask, state_evo, rng_mode));
}
};
struct RNNDescriptor
: public Descriptor<miopenRNNDescriptor,
@ -141,11 +131,6 @@ struct RNNDescriptor
miopenRNNBiasMode_t bias_mode, miopenRNNAlgo_t algorithm, miopenDataType_t datatype) {
MIOPEN_CHECK(miopenSetRNNDescriptor(mut_desc(), hidden_size, num_layers, input_mode, direction, rnn_mode, bias_mode, algorithm, datatype));
}
void setWithDropout(DropoutDescriptor& dropout_desc, int64_t hidden_size, int64_t num_layers, miopenRNNInputMode_t input_mode, miopenRNNDirectionMode_t direction,
miopenRNNMode_t rnn_mode, miopenRNNBiasMode_t bias_mode, miopenRNNAlgo_t algorithm, miopenDataType_t datatype) {
MIOPEN_CHECK(miopenSetRNNDescriptor_V2(mut_desc(), hidden_size, num_layers, dropout_desc.mut_desc(), input_mode, direction, rnn_mode, bias_mode, algorithm, datatype));
}
};
union Constant

73
aten/src/ATen/native/miopen/RNN_miopen.cpp
View File

@ -57,10 +57,6 @@ namespace at::native {
#include <ATen/TensorUtils.h>
#include <c10/hip/HIPCachingAllocator.h>
#include <rocrand/rocrand_xorwow.h>
#include <functional>
#include <iterator>
#include <sstream>
@ -72,36 +68,10 @@ namespace at::native {
namespace at { namespace native {
// Workspace copied from Conv_miopen.cpp but is put here inside anonymous namespace
// to avoid duplicate symbols and to avoid the need to expose as a public struct.
namespace {
struct Workspace {
Workspace(size_t size) : size(size), data(NULL) {
data = c10::hip::HIPCachingAllocator::raw_alloc(size);
}
Workspace(const Workspace&) = delete;
Workspace(Workspace&&) = default;
Workspace& operator=(Workspace&&) = default;
~Workspace() {
if (data) {
c10::hip::HIPCachingAllocator::raw_delete(data);
}
}
size_t size;
void* data;
};
} // anonymous
//RNNDescriptor.
struct RNNDescriptorParams {
int64_t hidden_size;
int64_t num_layers;
double dropout_rate;
uint64_t dropout_seed;
miopenRNNDirectionMode_t direction;
miopenRNNMode_t rnn_mode;
miopenDataType_t datatype;
@ -144,16 +114,6 @@ struct RNNDescriptorParams {
}
}
void set_dropout(double dropout_rate, uint64_t dropout_seed = 0) {
this->dropout_rate = dropout_rate;
if (dropout_seed == 0) {
// rand() returns 32 bit values so we combine two of them
this->dropout_seed = rand() << 32 | rand();
} else {
this->dropout_seed = dropout_seed;
}
}
void set(int64_t mode, int64_t hidden_size, int64_t num_layers, bool bidirectional, miopenDataType_t datatype, miopenRNNBiasMode_t bias_mode) {
this->set_mode(mode);
this->hidden_size = hidden_size;
@ -168,12 +128,6 @@ struct RNNDescriptorParams {
rnn_desc.set(hidden_size, num_layers, input_mode, direction, rnn_mode, bias_mode, algo, datatype);
return rnn_desc;
}
RNNDescriptor descriptorWithDropout(DropoutDescriptor& dropout_desc) const {
RNNDescriptor rnn_desc;
rnn_desc.setWithDropout(dropout_desc, hidden_size, num_layers, input_mode, direction, rnn_mode, bias_mode, algo, datatype);
return rnn_desc;
}
};
//TensorDescriptor list.
@ -250,8 +204,6 @@ struct RNNParams {
struct RNNDescriptors {
RNNDescriptor rnn_desc;
DropoutDescriptor dropout_desc;
std::unique_ptr<Workspace> dropout_states;
std::vector<TensorDescriptor> x_descs;
std::vector<TensorDescriptor> y_descs;
TensorDescriptor hx_desc;
@ -260,25 +212,7 @@ struct RNNDescriptors {
TensorDescriptor cy_desc;
RNNDescriptors(const RNNParams& fn, miopenHandle_t handle, Tensor x, Tensor y, Tensor hx, Tensor cx) {
if (fn.rnn.dropout_rate == 0.0) {
rnn_desc = fn.rnn.descriptor();
} else {
size_t statesSizeInBytes = 0;
miopenDropoutGetStatesSize(handle, &statesSizeInBytes);
size_t states_size = statesSizeInBytes / sizeof(rocrand_state_xorwow);
dropout_states = std::unique_ptr<Workspace>(new Workspace(states_size * sizeof(rocrand_state_xorwow)));
dropout_desc.set(handle,
fn.rnn.dropout_rate,
dropout_states->data,
dropout_states->size,
fn.rnn.dropout_seed,
false,
false,
miopenRNGType_t::MIOPEN_RNG_PSEUDO_XORWOW);
rnn_desc = fn.rnn.descriptorWithDropout(dropout_desc);
}
x_descs = fn.tensors.descriptors(x);
y_descs = fn.tensors.descriptors(y);
hx_desc.set(hx, 5);
@ -558,7 +492,7 @@ std::tuple<Tensor, Tensor, Tensor, Tensor, Tensor> miopen_rnn(
auto handle = getMiopenHandle();
miopenRNNAlgo_t algo = miopenRNNdefault;
fn.rnn.set_algo(algo);
fn.rnn.set_dropout(fn_dropout);
RNNDescriptors descs(fn, handle, x, y, hx, cx);
FilterDescriptor w_desc;
@ -617,6 +551,7 @@ std::tuple<Tensor, Tensor, Tensor, Tensor, Tensor> miopen_rnn(
}
return std::make_tuple(output, hy, cy, reserve, weight_buf);
}
std::tuple<Tensor, Tensor, Tensor, Tensor> miopen_rnn_backward_input(
@ -691,7 +626,6 @@ std::tuple<Tensor, Tensor, Tensor, Tensor> miopen_rnn_backward_input(
miopenRNNAlgo_t algo = miopenRNNdefault;
fn.rnn.set_algo(algo);
fn.rnn.set_dropout(fn_dropout);
RNNDescriptors descs(fn, handle, x, y, hx, cx);
FilterDescriptor w_desc;
@ -786,7 +720,6 @@ std::vector<Tensor> miopen_rnn_backward_weight(
miopenRNNAlgo_t algo = miopenRNNdefault;
fn.rnn.set_algo(algo);
fn.rnn.set_dropout(fn_dropout);
RNNDescriptors descs(fn, handle, x, y, hx, cx);
FilterDescriptor w_desc;
@ -976,6 +909,6 @@ REGISTER_CUDA_DISPATCH(lstm_miopen_stub, &lstm_miopen)
REGISTER_CUDA_DISPATCH(lstm_packed_miopen_stub, &lstm_packed_miopen)
} // anonymous namespace
}} // namespace native
}} //namespace native.
#endif