forked from Xilinx/finn-hlslib
-
Notifications
You must be signed in to change notification settings - Fork 0
/
convlayer.h
executable file
·202 lines (187 loc) · 9.91 KB
/
convlayer.h
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
/******************************************************************************
* Copyright (c) 2019, Xilinx, Inc.
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright notice,
* this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* 3. Neither the name of the copyright holder nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO,
* THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
* EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
* PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS;
* OR BUSINESS INTERRUPTION). HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,
* WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF
* ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
*****************************************************************************/
/******************************************************************************
*
* Authors: Giulio Gambardella <[email protected]>
* Thomas B. Preusser <[email protected]>
* Marie-Curie Fellow, Xilinx Ireland, Grant Agreement No. 751339
* Christoph Doehring <[email protected]>
*
* \file convlayer.h
*
* Library of templated HLS functions for BNN deployment.
* This file lists a set of convenience functions used to implement
* convolutional layers.
*
*****************************************************************************/
#ifndef CONVLAYER_H
#define CONVLAYER_H
#include <ap_int.h>
#include <hls_stream.h>
#include "streamtools.h"
#include "mvau.hpp"
/**
* \brief Convolutional layer implementation
*
* The function implements a generic convolutional layer, and it's basically composed of the sliding window generator
* implemeting the im2col algorithm and the Matrix_Vector_Activate_Batch function to perform computation.
*
* \tparam ConvKernelDim Dimension of the convolutional kernel (assumed square)
* \tparam IFMChannels Number of Input Feature Maps
* \tparam IFMDim Width and Height of the Input Feature Map (assumed square)
* \tparam OFMChannels Number of Output Feature Maps
* \tparam OFMDim Width and Height of the Output Feature Map (assumed square)
* \tparam SIMD Number of input columns computed in parallel
* \tparam PE Number of output rows computed in parallel
* \tparam TSrcI DataType of the input activation (as used in the MAC)
* \tparam TDstI DataType of the output activation (as generated by the activation)
* \tparam TWeightI DataType of the weights (as used in the MAC)
* \tparam InStreamW Width of the input stream
* \tparam OutStreamW Width of the output stream
* \tparam TW DataType of the weights matrix - safely deducible from the paramaters
* \tparam TA DataType of the activation class (e.g. thresholds) - safely deducible from the paramaters
* \tparam R DataType for the resource used for FPGA implementation of the MAC - safely deducible from the paramaters
*
* \param in Input stream
* \param out Output stream
* \param weights Weights matrix (currently supports BinaryWeights or FixedPointWeights)
* \param activation Activation class
* \param reps Number of time the function has to be repeatedly executed (e.g. number of images)
* \param r Resource type for the hardware implementation of the MAC block
*/
template<
unsigned int ConvKernelDim,
unsigned int IFMChannels,
unsigned int IFMDim,
unsigned int OFMChannels,
unsigned int OFMDim,
unsigned int SIMD, // number of SIMD lanes
unsigned int PE, // number of PEs
typename TSrcI = Identity, // redefine I/O interpretation as needed for input activations
typename TDstI = Identity, // redefine I/O interpretation as needed for output activations
typename TWeightI = Identity, // redefine I/O interpretation as needed for weigths
int InStreamW, int OutStreamW, // safely deducible (stream width must be int though!)
typename TW, typename TA, typename R
>
void ConvLayer_Batch(hls::stream<ap_uint<InStreamW>> &in,
hls::stream<ap_uint<OutStreamW>> &out,
TW const &weights,
TA const &activation,
unsigned const reps,
R const &r) {
#pragma HLS INLINE
unsigned const MatrixW = ConvKernelDim * ConvKernelDim * IFMChannels;
unsigned const MatrixH = OFMChannels;
unsigned const InpPerImage = IFMDim*IFMDim*IFMChannels/InStreamW * TSrcI::width;
WidthAdjustedInputStream <InStreamW, SIMD*TSrcI::width, InpPerImage> wa_in (in, reps);
WidthAdjustedOutputStream <PE*TDstI::width, OutStreamW, OFMDim * OFMDim * (OFMChannels / PE)> mvOut (out, reps);
hls::stream<ap_uint<SIMD*TSrcI::width> > convInp("StreamingConvLayer_Batch.convInp");
ConvolutionInputGenerator<ConvKernelDim, IFMChannels, TSrcI::width, IFMDim,
OFMDim, SIMD,1>(wa_in, convInp, reps, ap_resource_dflt());
Matrix_Vector_Activate_Batch<MatrixW, MatrixH, SIMD, PE, 1, TSrcI, TDstI, TWeightI>
(static_cast<hls::stream<ap_uint<SIMD*TSrcI::width>>&>(convInp),
static_cast<hls::stream<ap_uint<PE*TDstI::width>>&> (mvOut),
weights, activation, reps* OFMDim * OFMDim, r);
}
/**
* \brief Convolutional layer implementation
*
* The function implements a generic convolutional layer, and it's basically composed of the sliding window generator
* implemeting the im2col algorithm and the Matrix_Vector_Activate_Batch function to perform computation.
*
* \tparam ConvKernelDim Dimension of the convolutional kernel (assumed square)
* \tparam IFMChannels Number of Input Feature Maps
* \tparam IFMDim Width and Height of the Input Feature Map (assumed square)
* \tparam OFMChannels Number of Output Feature Maps
* \tparam OFMDim Width and Height of the Output Feature Map (assumed square)
* \tparam STRIDE Stride of the convolutional kernel
*
* \tparam SIMD Number of input columns computed in parallel
* \tparam PE Number of output rows computed in parallel
* \tparam MMV Number of output pixels computed in parallel
*
* \tparam TSrcI DataType of the input activation (as used in the MAC)
* \tparam TDstI DataType of the output activation (as generated by the activation)
* \tparam TWeightI DataType of the weights (as used in the MAC)
* \tparam InStreamW Width of the input stream
* \tparam OutStreamW Width of the output stream
* \tparam TW DataType of the weights matrix - safely deducible from the paramaters
* \tparam TA DataType of the activation class (e.g. thresholds) - safely deducible from the paramaters
* \tparam R DataType for the resource used for FPGA implementation of the MAC - safely deducible from the paramaters
*
* \param in Input stream
* \param out Output stream
* \param weights Weights matrix (currently supports BinaryWeights or FixedPointWeights)
* \param activation Activation class
* \param reps Number of time the function has to be repeatedly executed (e.g. number of images)
* \param r Resource type for the hardware implementation of the MAC block
*/
template<
unsigned int ConvKernelDim,
unsigned int IFMChannels,
unsigned int IFMDim,
unsigned int OFMChannels,
unsigned int OFMDim,
unsigned int STRIDE,
unsigned int SIMD, // number of SIMD lanes
unsigned int PE, // number of PEs
unsigned int MMV,
typename TSrcI = Identity, // redefine I/O interpretation as needed for input activations
typename TDstI = Identity, // redefine I/O interpretation as needed for output activations
typename TWeightI = Identity, // redefine I/O interpretation as needed for weigths
int InStreamW, int OutStreamW, // safely deducible (stream width must be int though!)
typename TW, typename TA, typename R
>
void ConvLayer_Batch_MMV(hls::stream<ap_uint<InStreamW>> &in,
hls::stream<ap_uint<OutStreamW>> &out,
TW const &weights,
TA const &activation,
unsigned const reps,
R const &r) {
#pragma HLS INLINE
unsigned const MatrixW = ConvKernelDim * ConvKernelDim * IFMChannels;
unsigned const MatrixH = OFMChannels;
unsigned const InpPerImage = IFMDim*IFMDim*IFMChannels/InStreamW * TSrcI::width;
const unsigned int mmvReps = (reps * OFMDim * OFMDim) / MMV;
WidthAdjustedInputStream <InStreamW, SIMD*TSrcI::width, InpPerImage> wa_in (in, reps);
WidthAdjustedOutputStream <PE*TDstI::width*MMV, OutStreamW, OFMDim * OFMDim * (OFMChannels / PE)/MMV> mvOut (out, reps);
stream<MultiChanData<MMV, PE * TDstI::width> > mmv2dwc("mmv2dwc");
hls::stream<MultiChanData<MMV, SIMD *TSrcI::width> > convInp("StreamingConvLayer_Batch.convInp");
ConvolutionInputGenerator_MMV<ConvKernelDim, IFMChannels, TSrcI::width, IFMDim,
OFMDim, SIMD, STRIDE, MMV>(wa_in, convInp, reps, ap_resource_dflt());
Matrix_Vector_Activate_Batch<MatrixW, MatrixH, SIMD, PE, MMV, TSrcI, TDstI, TWeightI>
(static_cast<hls::stream<MultiChanData<MMV,SIMD*TSrcI::width>>&>(convInp),
static_cast<hls::stream<MultiChanData<MMV,PE*TDstI::width>>&>(mmv2dwc),
weights, activation, mmvReps, r);
FlattenMultiChanData<MMV, PE * TDstI::width>(mmv2dwc, mvOut, mmvReps); // I guess it should be flattened after the MMV DWC
}
#endif