61 lines
3.1 KiB
Plaintext
61 lines
3.1 KiB
Plaintext
/// Copyright (c) 2020 ARM Limited.
|
|
///
|
|
/// SPDX-License-Identifier: MIT
|
|
///
|
|
/// Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
/// of this software and associated documentation files (the "Software"), to deal
|
|
/// in the Software without restriction, including without limitation the rights
|
|
/// to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
/// copies of the Software, and to permit persons to whom the Software is
|
|
/// furnished to do so, subject to the following conditions:
|
|
///
|
|
/// The above copyright notice and this permission notice shall be included in all
|
|
/// copies or substantial portions of the Software.
|
|
///
|
|
/// THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
/// IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
/// FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
/// AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
/// LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
/// OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
/// SOFTWARE.
|
|
///
|
|
|
|
namespace armnn
|
|
{
|
|
/**
|
|
@page converter_and_quantizer The ArmNN Converter and Quantizer
|
|
@tableofcontents
|
|
|
|
@section S10_converter The ArmNN Converter
|
|
|
|
The `ArmnnConverter` is a program for converting neural networks from other formats to Arm NN format.
|
|
Currently the program supports models in Caffe, Onnx, Tensorflow Protocol Buffers and Tensorflow Lite FlatBuffers formats. Run the program with no arguments to see command-line help.
|
|
|
|
For more information about the layers that are supported, see <a href="parsers.xhtml">parsers</a>.
|
|
<br/><br/><br/><br/>
|
|
|
|
@section S11_quantizer The ArmNN Quantizer
|
|
|
|
The `ArmnnQuantizer` is a program for loading a 32-bit float network into ArmNN and converting it into a quantized asymmetric 8-bit or quantized symmetric 16-bit network.
|
|
It supports static quantization by default, dynamic quantization is enabled if CSV file of raw input tensors is provided. Run the program with no arguments to see command-line help.
|
|
|
|
|
|
|Cmd:|||
|
|
| ---|---|---|
|
|
| -h | --help | Display help messages |
|
|
| -f | --infile | Input file containing float 32 ArmNN Input Graph |
|
|
| -s | --scheme | Quantization scheme, "QAsymm8" or "QSymm16". Default value: QAsymm8 |
|
|
| -c | --csvfile | CSV file containing paths for raw input tensors for dynamic quantization. If unset, static quantization is used |
|
|
| -p | --preserve-data-type | Preserve the input and output data types. If unset, input and output data types are not preserved |
|
|
| -d | --outdir | Directory that output file will be written to |
|
|
| -o | --outfile | ArmNN output file name |
|
|
|
|
<br/>
|
|
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.sh
|
|
./ArmnnQuantizer -f /path/to/armnn/input/graph/ -s "QSymm16" -c /path/to/csv/file -p 1 -d /path/to/output -o outputFileName
|
|
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
|
|
**/
|
|
}
|