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
|
// Copyright 2004-present Facebook. All Rights Reserved.
#include "caffe2/quantization/server/int8_gen_quant_params.h"
#include <functional>
namespace caffe2 {
using namespace std;
using namespace dnnlowp;
// Expilictly register TypeMeta
CAFFE_KNOWN_TYPE(unique_ptr<Int8QuantSchemeBlob>);
CAFFE_KNOWN_TYPE(unique_ptr<Int8QuantParamsBlob>);
REGISTER_CPU_OPERATOR(
Int8GenQuantParams,
Int8GenQuantParamsOp<CPUContext, DefaultEngine>);
OPERATOR_SCHEMA(Int8GenQuantParams)
.NumInputs(2)
.NumOutputs(1)
.TensorInferenceFunction([](const OperatorDef& /* def */,
const vector<TensorShape>& in) {
vector<TensorShape> out(1);
out[0].set_data_type(TensorProto_DataType_FLOAT);
out[0].add_dims(1);
return out;
})
.Input(
0,
"X",
"The input data, or last N samples of the output activations.")
.Input(
1,
"quant_scheme",
"Int8QuantSchemeBlob that specifies the quantization kind and preserve_sparsity options when generating the quant params.")
.Output(
0,
"quant_param",
"Int8QuantParamsBlob that contains the scale and zero_point info in TensorQuantizationParams type.")
.SetDoc(
R"DOC(Operator wrapper for generating int8 tensor quantization parameters given the input data and quant scheme)DOC");
} // namespace caffe2
|
