Skip to main content

How to delete the parameter of a Neural Network?


Mathematica Neural Networks based on MXNet. And there is lots of model in MXNet(MXNet Model Zoo).


For example,I follow a tutorial An introduction to the MXNet API — part 4, It uses Inception v3(net using MXNet).


I can use this code to load this pretrained model


<< NeuralNetworks`;
net = ImportMXNetModel["F:\\model\\Inception-BN-symbol.json",
"F:\\model\\Inception-BN-0126.params"];

labels = Import["http://data.dmlc.ml/models/imagenet/synset.txt", "List"];
incepNet = NetChain[{ElementwiseLayer[#*255 &], net},
"Input" -> NetEncoder[{"Image", {224, 224}, ColorSpace -> "RGB"}],
"Output" -> NetDecoder[{"Class", labels}]]

enter image description here


 TakeLargestBy[incepNet[Import["https://i.stack.imgur.com/893F0.jpg"],
"Probabilities"], Identity, 5]//Dataset

enter image description here



Dynamic@Column[{#, incepNet[#]} &@CurrentImage[]]

enter image description here


FileByteCount[Export["Inception v3.wlnet", incepNet]]/2^20.
(*44.0188 MB*)

I want to share this model to others, but it's too big (about 44 MB) because there are lots of parameters in this pretrained model.


If I only want to share this model structure to others,I think deNetInitialize is necessary.


enter image description here


You can see UninitializedEvaluationNet only need very little memory.



In order to share nets like UninitializedEvaluationNet, can we define a deNetInitialize function?



Answer



The weights in the neural network are stored in the form of RawArray. This can be seen by comparing their full form


NetInitialize@ConvolutionLayer[1, {2}, "Input" -> {1, 3}] // InputForm
(* ConvolutionLayer[<|"Type" -> "Convolution",
"Arrays" -> <|"Weights" -> RawArray["Real32", {{{0.4258739650249481, -0.9216004014015198}}}],
"Biases" -> RawArray["Real32", {0.}]|>, "Parameters" -> <|"OutputChannels" -> 1, "KernelSize" -> {2}, "Stride" -> {1},
"PaddingSize" -> {0}, "Dilation" -> {1}, "Dimensionality" -> 1, "InputChannels" -> 1, "$GroupNumber" -> 1, "$InputSize" -> {3},
"$OutputSize" -> {2}|>, "Inputs" -> <|"Input" -> TensorT[{1, 3}, RealT]|>, "Outputs" -> <|"Output" -> TensorT[{1, 2}, RealT]|>|>,
<|"Version" -> "11.1.1"|>] *)


ConvolutionLayer[1, {2}, "Input" -> {1, 3}] // InputForm
(* ConvolutionLayer[<|"Type" -> "Convolution", "Arrays" -> <|"Weights" -> TensorT[{1, 1, 2}, RealT],
"Biases" -> Nullable[TensorT[{1}, RealT]]|>, "Parameters" -> <|"OutputChannels" -> 1, "KernelSize" -> {2}, "Stride" -> {1},
"PaddingSize" -> {0}, "Dilation" -> {1}, "Dimensionality" -> 1, "InputChannels" -> 1, "$GroupNumber" -> 1, "$InputSize" -> {3},
"$OutputSize" -> {2}|>, "Inputs" -> <|"Input" -> TensorT[{1, 3}, RealT]|>, "Outputs" -> <|"Output" -> TensorT[{1, 2}, RealT]|>|>,
<|"Version" -> "11.1.1"|>] *)

To remove the weights, we only need to replace the RawArray with the corresponding place holders. Here is a solution using string manipulation


netUnInitialize[net_] := 

Activate@Replace[
ToExpression[
StringReplace[
ToString[InputForm[net]], {"NetChain" -> "Inactive[NetChain]",
"NetGraph" -> "Inactive[NetGraph]",
"RawArray" -> "Inactive[RawArray]"}]],
Inactive[RawArray][x_, y_] :>
NeuralNetworks`TensorT[Dimensions[y],
NeuralNetworks`RealT], ∞]


FileByteCount[Export["~/Downloads/incept.wlnet",netUnInitialize[incepNet]]]
(* 40877 *)

Comments

Popular posts from this blog

plotting - Plot 4D data with color as 4th dimension

I have a list of 4D data (x position, y position, amplitude, wavelength). I want to plot x, y, and amplitude on a 3D plot and have the color of the points correspond to the wavelength. I have seen many examples using functions to define color but my wavelength cannot be expressed by an analytic function. Is there a simple way to do this? Answer Here a another possible way to visualize 4D data: data = Flatten[Table[{x, y, x^2 + y^2, Sin[x - y]}, {x, -Pi, Pi,Pi/10}, {y,-Pi,Pi, Pi/10}], 1]; You can use the function Point along with VertexColors . Now the points are places using the first three elements and the color is determined by the fourth. In this case I used Hue, but you can use whatever you prefer. Graphics3D[ Point[data[[All, 1 ;; 3]], VertexColors -> Hue /@ data[[All, 4]]], Axes -> True, BoxRatios -> {1, 1, 1/GoldenRatio}]

plotting - Filling between two spheres in SphericalPlot3D

Manipulate[ SphericalPlot3D[{1, 2 - n}, {θ, 0, Pi}, {ϕ, 0, 1.5 Pi}, Mesh -> None, PlotPoints -> 15, PlotRange -> {-2.2, 2.2}], {n, 0, 1}] I cant' seem to be able to make a filling between two spheres. I've already tried the obvious Filling -> {1 -> {2}} but Mathematica doesn't seem to like that option. Is there any easy way around this or ... Answer There is no built-in filling in SphericalPlot3D . One option is to use ParametricPlot3D to draw the surfaces between the two shells: Manipulate[ Show[SphericalPlot3D[{1, 2 - n}, {θ, 0, Pi}, {ϕ, 0, 1.5 Pi}, PlotPoints -> 15, PlotRange -> {-2.2, 2.2}], ParametricPlot3D[{ r {Sin[t] Cos[1.5 Pi], Sin[t] Sin[1.5 Pi], Cos[t]}, r {Sin[t] Cos[0 Pi], Sin[t] Sin[0 Pi], Cos[t]}}, {r, 1, 2 - n}, {t, 0, Pi}, PlotStyle -> Yellow, Mesh -> {2, 15}]], {n, 0, 1}]

plotting - Mathematica: 3D plot based on combined 2D graphs

I have several sigmoidal fits to 3 different datasets, with mean fit predictions plus the 95% confidence limits (not symmetrical around the mean) and the actual data. I would now like to show these different 2D plots projected in 3D as in but then using proper perspective. In the link here they give some solutions to combine the plots using isometric perspective, but I would like to use proper 3 point perspective. Any thoughts? Also any way to show the mean points per time point for each series plus or minus the standard error on the mean would be cool too, either using points+vertical bars, or using spheres plus tubes. Below are some test data and the fit function I am using. Note that I am working on a logit(proportion) scale and that the final vertical scale is Log10(percentage). (* some test data *) data = Table[Null, {i, 4}]; data[[1]] = {{1, -5.8}, {2, -5.4}, {3, -0.8}, {4, -0.2}, {5, 4.6}, {1, -6.4}, {2, -5.6}, {3, -0.7}, {4, 0.04}, {5, 1.0}, {1, -6.8}, {2, -4.7}, {3, -1....