.. blockdiag::
blockdiag {
default_fontsize=8
span_width = 32;
span_height = 20;
NNabla1 [label = "NNabla", color="lime", shape="roundedbox", width=80, height=60, fontsize=12];
NNabla2 [label = "Use NNabla as\nRuntime", color="lime", shape="roundedbox", width=80, height=60];
Other [label = "Other\n(Caffe2 etc.)", shape="roundedbox", width=80, height=60];
ONNX1 [label = "ONNX", color="mediumslateblue", width=40, height=20];
ONNX2 [label = "ONNX", color="mediumslateblue", width=40, height=20];
Conv1 [label = "File Format\nConverter", color="lime", shape="roundedbox", width=80, height=60, fontsize=10];
Conv2 [label = "File Format\nConverter", color="lime", shape="roundedbox", width=80, height=60, fontsize=10];
NNP1 [label = "NNP", color="cyan", width=40, height=20];
NNP2 [label = "NNP", color="cyan", width=40, height=20];
NNB [label = "NNB", color="cyan", width=40, height=20];
CSRC [label = "C Source\ncode", color="seagreen", width=40];
OtherRuntime [label = "Other runtime", shape="roundedbox", width=80];
NNablaCRuntime [label = "NNabla C\nRuntime", color="lime", shape="roundedbox", width=80];
Product [label = "Implement to\nproduct", shape="roundedbox", width=80, height=60];
NNabla1 -> NNP1;
Other -> ONNX1 -> Conv1 -> NNP1;
NNP1 -> Conv2;
Conv2 -> ONNX2 -> OtherRuntime;
Conv2 -> NNB -> NNablaCRuntime;
Conv2 -> CSRC -> Product;
Conv2 -> NNP2 -> NNabla2;
}
File format converter will realize Neural Network Libraries (or Console) workflow with ONNX file format, and also NNabla C Runtime.
File format converter has following functions.
- Convert NNP variations to valid NNP
- Convert ONNX to NNP
- Convert NNP to ONNX
- Convert NNP to NNB(Binary format for NNabla C Runtime)
- Experimental: Convert NNP to C Source code for NNabla C Runtime
IMPORTANT NOTICE: This file format converter still has some known problems.
- Supported ONNX operator is limited. See :any:`onnx/operator_coverage`.
- Converting NNP to C Source code is still experimental. It should work but did not tested well.
.. blockdiag::
blockdiag {
default_group_color = white;
INPUT [label="<<file>>\nINPUT", color="lime"];
OUTPUT [label="<<file>>\nOUTPUT", color="lime"];
PROCESS [label="Process\n(Split, Expand, etc.)", shape="roundedbox"];
proto [label="proto", color="cyan", width=60, height=20];
INPUT -> proto [label="import"];
group {
orientation = portrait;
proto <-> PROCESS;
}
proto -> OUTPUT [label="export"];
}
This file format converter uses protobuf defined in Neural Network Libraries as intermediate format.
While this is not a generic file format converter, this is the specified converter for Neural Network Libraries.
This converter can specify both inputs and outputs for ONNX file, but if ONNX file contains a function unsupported by Neural Network Libraries, it may cause error in conversion.
This converter also provides some intermediate process functionalities. See :ref:Process.
NNP is file format of NNabla.
NNP format is described at :any:`../../format`.
But with this file format converter is work with several variation of NNP.
- Standard NNP format (.nnp)
- Contents of NNP files(.nntxt, .prototxt, .h5, .protobuf)
- Training is not supported
- Only supports operator set 3
- Not all functions are supported. See :any:`onnx/operator_coverage`.
- Only limited Neural Network Console projects supported. See :any:`onnx/neural_network_console_example_coverage`.
- In some case you must install onnx package by hand. For example you can install with command pip install onnx or if you want to install system wide, you can install with command sudo -HE pip install onnx.
NNB is compact binary format for NNabla C Runtime. It is designed for nnabla-c-runtime.
File format converter supports C source code output for nnabla-c-runtime.
Neural Network Console supports LoopControl pseudo functions RepeatStart, RepeatEnd, RecurrentInput, RecurrentOutput or Delay.
Currently, these functions are not supported by Neural Network Libraries directly.
The file format converter expands the network and removes these pseudo functions by default.
If you want to preserve these, specify command line option --nnp-no-expand-network when converting files.
You can split network with --split option.
See :ref:`Splitting network` to use this functionality.
Sometimes we need convert NNP to NNP.
Most major usecase, expand repeat or recurrent network supported by Neural Network Console but does not supported by C++ API.
$ nnabla_cli convert --nnp-no-expand-network input.nnp output.nnp
Current version of Neural Network Console outputs .nntxt and .h5 as training result.
Then we need to convert separated files into single NNP and parameters store with protobuf format.
$ nnabla_cli convert net.nntxt parameters.h5 output.nnp
$ nnabla_cli convert --nnp-no-expand-network net.nntxt parameters.h5 output.nnp
$ nnabla_cli convert --nnp-no-expand-network --nnp-parameter-h5 net.nntxt parameters.h5 output.nnp
$ nnabla_cli convert --nnp-parameter-nntxt net.nntxt parameters.h5 output.nntxt
$ nnabla_cli convert input.nnp output.onnx
$ nnabla_cli convert input.onnx output.nnp
$ nnabla_cli convert input.nnp output.nnb
$ nnabla_cli convert -O CSRC input.onnx output-dir
Splitting network is a bit complicated and can be troublesome.
NNP file could have multiple Executor networks, but Split supports only single network to split.
First, you must confirm how many Executors there are in the NNP, and specify what executor to split with nnabla_cli dump.
$ nnabla_cli dump squeezenet11.files/SqueezeNet-1.1/*.{nntxt,h5}
2018-08-27 15:02:40,006 [nnabla][INFO]: Initializing CPU extension...
Importing squeezenet11.files/SqueezeNet-1.1/net.nntxt
Importing squeezenet11.files/SqueezeNet-1.1/parameters.h5
Expanding Training.
Expanding Top5Error.
Expanding Top1Error.
Expanding Runtime.
Optimizer[0]: Optimizer
Optimizer[0]: (In) Data variable[0]: Name:TrainingInput Shape:[-1, 3, 480, 480]
Optimizer[0]: (In) Data variable[1]: Name:SoftmaxCrossEntropy_T Shape:[-1, 1]
Optimizer[0]: (Out)Loss variable[0]: Name:SoftmaxCrossEntropy Shape:[-1, 1]
Monitor [0]: train_error
Monitor [0]: (In) Data variable[0]: Name:Input Shape:[-1, 3, 320, 320]
Monitor [0]: (In) Data variable[1]: Name:Top5Error_T Shape:[-1, 1]
Monitor [0]: (Out)Monitor variable[0]: Name:Top5Error Shape:[-1, 1]
Monitor [1]: valid_error
Monitor [1]: (In) Data variable[0]: Name:Input Shape:[-1, 3, 320, 320]
Monitor [1]: (In) Data variable[1]: Name:Top1rror_T Shape:[-1, 1]
Monitor [1]: (Out)Monitor variable[0]: Name:Top1rror Shape:[-1, 1]
Executor [0]: Executor
Executor [0]: (In) Data variable[0]: Name:Input Shape:[-1, 3, 320, 320]
Executor [0]: (Out)Output variable[0]: Name:y' Shape:[-1, 1000]
As above output now you know only 1 executor.
Then you can show executor information with nnabla_cli dump -E0.
$ nnabla_cli dump -E0 squeezenet11.files/SqueezeNet-1.1/*.{nntxt,h5}
2018-08-27 15:03:26,547 [nnabla][INFO]: Initializing CPU extension...
Importing squeezenet11.files/SqueezeNet-1.1/net.nntxt
Importing squeezenet11.files/SqueezeNet-1.1/parameters.h5
Try to leave only executor[Executor].
Expanding Runtime.
Executor [0]: Executor
Executor [0]: (In) Data variable[0]: Name:Input Shape:[-1, 3, 320, 320]
Executor [0]: (Out)Output variable[0]: Name:y' Shape:[-1, 1000]
You can get list of function adding -F option.
$ nnabla_cli dump -FE0 squeezenet11.files/SqueezeNet-1.1/*.{nntxt,h5}
2018-08-27 15:04:10,954 [nnabla][INFO]: Initializing CPU extension...
Importing squeezenet11.files/SqueezeNet-1.1/net.nntxt
Importing squeezenet11.files/SqueezeNet-1.1/parameters.h5
Try to leave only executor[Executor].
Expanding Runtime.
Executor [0]: Executor
Executor [0]: (In) Data variable[0]: Name:Input Shape:[-1, 3, 320, 320]
Executor [0]: (Out)Output variable[0]: Name:y' Shape:[-1, 1000]
Executor [0]: Function[ 0 ]: Type: Slice Name: Slice
Executor [0]: Function[ 1 ]: Type: ImageAugmentation Name: ImageAugmentation
Executor [0]: Function[ 2 ]: Type: MulScalar Name: SqueezeNet/MulScalar
Executor [0]: Function[ 3 ]: Type: AddScalar Name: SqueezeNet/AddScalar
Executor [0]: Function[ 4 ]: Type: Convolution Name: SqueezeNet/Convolution
Executor [0]: Function[ 5 ]: Type: ReLU Name: SqueezeNet/ReLU
Executor [0]: Function[ 6 ]: Type: MaxPooling Name: SqueezeNet/MaxPooling
SNIP...
Executor [0]: Function[ 63 ]: Type: ReLU Name: SqueezeNet/FireModule_8/Expand1x1ReLU
Executor [0]: Function[ 64 ]: Type: Concatenate Name: SqueezeNet/FireModule_8/Concatenate
Executor [0]: Function[ 65 ]: Type: Dropout Name: SqueezeNet/Dropout
Executor [0]: Function[ 66 ]: Type: Convolution Name: SqueezeNet/Convolution_2
Executor [0]: Function[ 67 ]: Type: ReLU Name: SqueezeNet/ReLU_2
Executor [0]: Function[ 68 ]: Type: AveragePooling Name: SqueezeNet/AveragePooling
Executor [0]: Function[ 69 ]: Type: Reshape Name: SqueezeNet/Reshape
Executor [0]: Function[ 70 ]: Type: Identity Name: y'
If you want to get network without Image Augmentation, according to above output, ImageAugmentation is placed on index 2. With splitting after index 3, you can get network without ImageAugmentation. You must specify -E0 -S 3- option to nnabla_cli convert This command rename output to XXX_S_E.nnp, XXX is original name, S is start function index, and E is end function index.
$ nnabla_cli convert -E0 -S 3- squeezenet11.files/SqueezeNet-1.1/*.{nntxt,h5} splitted.nnp
2018-08-27 15:20:21,950 [nnabla][INFO]: Initializing CPU extension...
Importing squeezenet11.files/SqueezeNet-1.1/net.nntxt
Importing squeezenet11.files/SqueezeNet-1.1/parameters.h5
Try to leave only executor[Executor].
Expanding Runtime.
Shrink 3 to 70.
Output to [splitted_3_70.nnp]
Finally you got splitted_3_70.nnp as splitted output. You can check splitted NNP with nnabla_cli dump
NOTE: Input shape is changed from original network. New input shape is same as start function's input.
$ nnabla_cli dump splitted_3_70.nnp
2018-08-27 15:20:28,021 [nnabla][INFO]: Initializing CPU extension...
Importing splitted_3_70.nnp
Expanding Runtime.
Executor [0]: Executor
Executor [0]: (In) Data variable[0]: Name:SqueezeNet/MulScalar Shape:[-1, 3, 227, 227]
Executor [0]: (Out)Output variable[0]: Name:y' Shape:[-1, 1000]
Done.