diff options
Diffstat (limited to 'docs/03_converter_quantizer.dox')
-rw-r--r-- | docs/03_converter_quantizer.dox | 60 |
1 files changed, 60 insertions, 0 deletions
diff --git a/docs/03_converter_quantizer.dox b/docs/03_converter_quantizer.dox new file mode 100644 index 0000000000..ebfacd473e --- /dev/null +++ b/docs/03_converter_quantizer.dox @@ -0,0 +1,60 @@ +/// Copyright (c) 2017 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 +~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + +**/ +} |