Skip to content
This repository has been archived by the owner on Nov 17, 2023. It is now read-only.

[MXNET-133] Model Quantization with Calibration #9552

Merged
merged 16 commits into from
Mar 26, 2018
Merged

[MXNET-133] Model Quantization with Calibration #9552

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
16 commits
Select commit Hold shift + click to select a range
432cac0
[Quantization] 8bit Quantization and GPU Support
ZihengJiang Apr 28, 2017
115a07d
[Quantization] Add calibration flow and refactor code
reminisce Sep 6, 2017
7590312
Fix compile error
reminisce Mar 12, 2018
790f29c
Fix CI
reminisce Mar 12, 2018
5c593de
Remove tests that should not run on P3
reminisce Mar 12, 2018
121e655
Remove unnecessary docker file
reminisce Mar 12, 2018
e100267
Fix registering quantized nn ops
reminisce Mar 13, 2018
fa6a419
Reformat Jenkinsfile and switch quantization to CUDA 9 (#9)
marcoabreu Mar 21, 2018
9f251ae
Address interface change cr
reminisce Mar 21, 2018
1995b73
Address comments and fix bugs
reminisce Mar 22, 2018
4f33f92
Make unit test stable
reminisce Mar 22, 2018
33964bd
Improve unit test
reminisce Mar 22, 2018
d9f2068
Address cr
reminisce Mar 23, 2018
0a943c3
Address cr
reminisce Mar 23, 2018
21512de
Fix flaky unit test layer_norm
reminisce Mar 23, 2018
7be4936
Fix doc
reminisce Mar 24, 2018
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
Prev Previous commit
Next Next commit
Fix registering quantized nn ops
  • Loading branch information
@reminisce
reminisce committed Mar 25, 2018
commit e10026768d56e82b7fd4d5799af1cbb7313ef154
8 changes: 4 additions & 4 deletions src/operator/quantization/quantized_conv.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,6 @@
* \brief
* \author Ziheng Jiang, Jun Wu
*/
#if MSHADOW_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
#include "../nn/convolution-inl.h"
#include "./quantization_utils.h"
#include "../tensor/matrix_op-inl.h"
Expand All @@ -50,6 +49,7 @@ struct QuantizedBiasAddKernel {
}
};

#if MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
template<typename SrcType, typename DstType, typename CmpType>
class QuantizedCuDNNConvOp {
public:
Expand Down Expand Up @@ -260,6 +260,7 @@ class QuantizedCuDNNConvOp {
float alpha_ = 1.0f;
float beta_ = 0.0f;
}; // class QuantizedCuDNNConvOp
#endif // MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6

void QuantizedConvForwardGPU(const nnvm::NodeAttrs& attrs,
const OpContext& ctx,
Expand All @@ -269,7 +270,7 @@ void QuantizedConvForwardGPU(const nnvm::NodeAttrs& attrs,
const ConvolutionParam& param = nnvm::get<ConvolutionParam>(attrs.parsed);
CHECK_EQ(param.kernel.ndim(), 2U)
<< "QuantizedConvForward<gpu> only supports 2D convolution for now";
#if MXNET_USE_CUDNN == 1
#if MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
typedef QuantizedCuDNNConvOp<int8_t, float, int32_t> QuantizedConvOpInt8;
#if DMLC_CXX11_THREAD_LOCAL
static thread_local QuantizedConvOpInt8 op;
Expand All @@ -280,12 +281,11 @@ void QuantizedConvForwardGPU(const nnvm::NodeAttrs& attrs,
op.Forward(ctx, inputs, req, outputs);
#else
LOG(FATAL) << "QuantizedConvForward<gpu> only supports cudnnConvolutionForward for now";
#endif // MXNET_USE_CUDNN
#endif // MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
}

NNVM_REGISTER_OP(_contrib_quantized_conv)
.set_attr<FCompute>("FCompute<gpu>", QuantizedConvForwardGPU);

} // namespace op
} // namespace mxnet
#endif
2 changes: 0 additions & 2 deletions src/operator/quantization/quantized_fully_connected.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -23,7 +23,6 @@
* \brief
* \author Ziheng Jiang, Jun Wu
*/
#if MSHADOW_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
#include "./quantization_utils.h"
#include "../mxnet_op.h"
#include "../nn/fully_connected-inl.h"
Expand Down Expand Up @@ -121,4 +120,3 @@ NNVM_REGISTER_OP(_contrib_quantized_fully_connected)

} // namespace op
} // namespace mxnet
#endif
8 changes: 4 additions & 4 deletions src/operator/quantization/quantized_pooling.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,6 @@
* Copyright (c) 2017 by Contributors
* \file quantized_pooling.cu
*/
#if MSHADOW_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
#include <mxnet/operator_util.h>
#include <vector>
#include "../nn/pooling-inl.h"
Expand All @@ -30,6 +29,7 @@
namespace mxnet {
namespace op {

#if MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
template<typename DType>
class QuantizedCuDNNPoolingOp {
public:
Expand Down Expand Up @@ -115,6 +115,7 @@ class QuantizedCuDNNPoolingOp {
cudnnTensorDescriptor_t out_desc_;
cudnnPoolingDescriptor_t pool_desc_;
}; // class QuantizedCuDNNPoolingOp
#endif // MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6

void QuantizedPoolingForwardGPU(const nnvm::NodeAttrs& attrs,
const OpContext& ctx,
Expand All @@ -124,7 +125,7 @@ void QuantizedPoolingForwardGPU(const nnvm::NodeAttrs& attrs,
const PoolingParam& param = nnvm::get<PoolingParam>(attrs.parsed);
CHECK_EQ(param.kernel.ndim(), 2U)
<< "QuantizedPoolingForward<gpu> only supports 2D convolution for now";
#if MXNET_USE_CUDNN == 1
#if MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
#if DMLC_CXX11_THREAD_LOCAL
static thread_local QuantizedCuDNNPoolingOp<int8_t> op;
#else
Expand All @@ -134,12 +135,11 @@ void QuantizedPoolingForwardGPU(const nnvm::NodeAttrs& attrs,
op.Forward(ctx.get_stream<gpu>(), inputs, req, outputs);
#else
LOG(FATAL) << "QuantizedPoolingForward<gpu> only supports cudnnPoolingForward for now";
#endif // MXNET_USE_CUDNN
#endif // MXNET_USE_CUDNN == 1 && CUDNN_MAJOR >= 6
}

NNVM_REGISTER_OP(_contrib_quantized_pooling)
.set_attr<FCompute>("FCompute<gpu>", QuantizedPoolingForwardGPU);

} // namespace op
} // namespace mxnet
#endif