Common tools for function fitting

This post describes the use of a set of common tools written in Python to generate and display synthetic datasets obtained by performing both scalar and vector generator functions of one or more variables or datasets obtained by approximating such functions through appropriate programs described in other posts on this website that deal with the fitting of mathematical objects (functions, curves and surfaces) both with machine learning algorithms and neural networks.

The areas affected by these tools are:

  • Creation of synthetic datasets
  • Comparative display of datasets
  • Diagnostics
The related code requires the version 3 of Python and also it requires also NumPy and MatPlotLib libraries.
To get the code of each tool follow the link at the beginning of each paragraph related to the tool itself; the paragraph Download of the complete code contains the link to the entire repository which deals with the approximation of mathematical functions with machine and deep learning techniques.

Creation of synthetic datasets

This chapter explains the tools for generating synthetic datasets from mathematical generator functions.


fx_gen.py

Goal of the fx_gen.py Python program is to generate a synthetic dataset in csv format starting from a real-valued and scalar generator function of a real-valued variable in a closed interval specifying the step of discretization.

The output csv file has two columns (with header): first column contains the sorted values of independent variable $x$ within the passed interval discretized by discretization step; second column contains the values of dependent variable, ie the values of function $f(x)$ correspondent to values of $x$ of first column.

To get the program usage you can run this following command: 

$ python fx_gen.py --help
and the output got is: 

usage: fx_gen.py [-h] [--version] --dsout DS_OUTPUT_FILENAME --funcx
                 FUNC_X_BODY [--xbegin RANGE_BEGIN] [--xend RANGE_END]
                 [--xstep RANGE_STEP] [--noise NOISE_BODY]

fx_gen.py generates a synthetic dataset file calling a one-variable real
function in an interval

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --dsout DS_OUTPUT_FILENAME
                        dataset output file (csv format)
  --funcx FUNC_X_BODY   f(x) body (lamba format)
  --xbegin RANGE_BEGIN  begin range (default:-5.0)
  --xend RANGE_END      end range (default:+5.0)
  --xstep RANGE_STEP    step range (default: 0.01)
  --noise NOISE_BODY    noise(sz) body (lamba format)
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --dsout: path (relative or absolute) of the csv file to be generated and got executing the dataset generator function.

  • --funcx: the generator function $y=f(x)$ of the time dataset in lamba body format where $x$ is the independent variable.

  • --xbegin e --xend: the closed interval of the variable $x$, between --xbegin and --xend.

  • --xstep: the discretization step of the independent variable $x$.

  • --noise: the noise generator functon lamba body format, where $sz$ is the number of elements of the discretized interval between --xbegin and --xend.

An example of using the fx_gen.py program

Suppose you want to approximate the function $$f(x)=\sqrt{|x|}$$ in the range $[-6.0,6.0]$ Keeping in mind that np is the alias of NumPy library, the translation of this function in lambda body Python syntax is: 

np.sqrt(np.abs(x))
To generate the training dataset, run the following command: 

$ python fx_gen.py \
  --dsout mytrain.csv \
  --funcx "np.sqrt(np.abs(x))" \
  --xbegin -6.0 \
  --xend 6.0 \
  --xstep 0.01


fxy_gen.py

Goal of the fxy_gen.py Python program is to generate a synthetic dataset in csv format starting from a real-valued and scalar generator function of two real-valued variables in a closed rectangle specifying the step of discretization.

The output csv file has three columns (with header): first two columns contain the sorted values of independent variables $x$ and $y$ within the passed intervals discretized by their own discretization steps; third column contains the values of dependent variable, ie the values of function $f(x,y)$ correspondent to values of $x$ and $y$ of first two columns.

To get the program usage you can run this following command: 

$ python fxy_gen.py --help
and the output got is: 

usage: fxy_gen.py [-h] [--version] --dsout DS_OUTPUT_FILENAME --funcxy
                  FUNC_XY_BODY [--xbegin RANGE_XBEGIN] [--xend RANGE_XEND]
                  [--ybegin RANGE_YBEGIN] [--yend RANGE_YEND]
                  [--xstep RANGE_XSTEP] [--ystep RANGE_YSTEP]
				  [--noise NOISE_BODY]

fxy_gen.py generates a synthetic dataset file calling a two-variables real
function on a rectangle

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --dsout DS_OUTPUT_FILENAME
                        dataset output file (csv format)
  --funcxy FUNC_XY_BODY
                        f(x, y) body (lamba format)
  --xbegin RANGE_XBEGIN
                        begin x range (default:-5.0)
  --xend RANGE_XEND     end x range (default:+5.0)
  --ybegin RANGE_YBEGIN
                        begin y range (default:-5.0)
  --yend RANGE_YEND     end y range (default:+5.0)
  --xstep RANGE_XSTEP   step range of x (default: 0.01)
  --ystep RANGE_YSTEP   step range of y (default: 0.01)
  --noise NOISE_BODY    noise(sz) body (lamba format)
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --dsout: path (relative or absolute) of the csv file to be generated and got executing the dataset generator function.

  • --funcxy: the generator function $y=f(x,y)$ of the time dataset in lamba body format where $x$ and $y$ are the independent variables.

  • --xbegin e --xend: the closed interval of the variable $x$, between --xbegin and --xend.

  • --ybegin e --yend: the closed interval of the variable $y$, between --ybegin and --yend.

  • --xstep: the discretization step of the independent variable $x$.

  • --ystep: the discretization step of the independent variable $y$.

  • --noise: the noise generator functon lamba body format, where $sz$ is the number of elements of the discretized rectagle between --xbegin and --xend for --ybegin and --yend. interval between --xbegin and --xend.

An example of using the fxy_gen.py program

Suppose you want to approximate in the rectagle $[-6.0,6.0]\times[-6.0,6.0]$ the following function $$\sin \sqrt{x^2 + y^2}$$, Keeping in mind that np is the alias of NumPy library, the translation of this function in lambda body Python syntax is: 

np.sin(np.sqrt(x**2 + y**2))
To generate the training dataset, run the following command: 

$ python fxy_gen.py \
  --dsout mytrain.csv \
  --funcxy "np.sin(np.sqrt(x**2 + y**2))" \
  --xbegin -6.0 \
  --xend 6.0 \
  --ybegin -6.0 \
  --yend 6.0 \
  --xstep 0.075 \
  --ystep 0.075


pmc2t_gen.py

Goal of the pmc2t_gen.py Python program is to generate a synthetic dataset in csv format starting from a couple of real-valued functions of one real-valued variable $t$, called parameter:
it takes in command line the two component functions to be approximated (in lambda body syntax), the interval of independent parameter $t$ (begin, end and discretization step) and it generates the dataset in an output csv file applying the two functions to the passed interval of the parameter $t$.
In fact the output csv file has three columns (with header): first column contains the sorted values of independent parameter $t$ within the passed interval discretized by discretization step; second and third columns contain respectively the values of the components $x$ and $y$, ie the values of functions $x(t)$ and $y(t)$ correspondent to values of $t$ of first column.
To get the program usage you can run this following command: 

$ python pmc2t_gen.py --help
and the output got is: 

usage: pmc2t_gen.py [-h] [--version] --dsout DS_OUTPUT_FILENAME --funcxt
                    FUNCX_T_BODY --funcyt FUNCY_T_BODY [--tbegin RANGE_BEGIN]
                    [--tend RANGE_END] [--tstep RANGE_STEP]
                    [--xnoise XNOISE_BODY] [--ynoise YNOISE_BODY]

pmc2t_gen.py generates a synthetic dataset file that contains the points of a
parametric curve on plane calling a couple of one-variable real functions in
an interval

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --dsout DS_OUTPUT_FILENAME
                        dataset output file (csv format)
  --funcxt FUNCX_T_BODY
                        x=x(t) body (lamba format)
  --funcyt FUNCY_T_BODY
                        y=y(t) body (lamba format)
  --tbegin RANGE_BEGIN  begin range (default:-5.0)
  --tend RANGE_END      end range (default:+5.0)
  --tstep RANGE_STEP    step range (default: 0.01)
  --xnoise XNOISE_BODY  noise(sz) body (lamba format) on x
  --ynoise YNOISE_BODY  noise(sz) body (lamba format) on y
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --dsout: path (relative or absolute) of the csv file to be generated and got executing the dataset generator function.

  • --funcxt: the generator function $x=fx(t)$ of the time dataset in lamba body format where $t$ is the independent variable.

  • --funcyt: the generator function $y=fy(t)$ of the time dataset in lamba body format where $t$ is the independent variable.

  • --tbegin e --tend: the closed interval of the variable $t$, between --tbegin and --tend.

  • --tstep: the discretization step of the independent variable $x$.

  • --xnoise: the noise generator functon on $x$ variable in lamba body format, where $sz$ is the number of elements of the discretized interval between --tbegin and --tend.

  • --ynoise: the noise generator functon on $y$ variable in lamba body format, where $sz$ is the number of elements of the discretized interval between --tbegin and --tend.

pmc3t_gen.py

Goal of the pmc3t_gen.py Python program is to generate a synthetic dataset in csv format starting from a triple of real-valued functions of one real-valued variable $t$, called parameter
it takes in command line the three component functions to be approximated (in lambda body syntax), the interval of independent parameter $t$ (begin, end and discretization step) and it generates the dataset in an output csv file applying the three functions to the passed interval of the parameter $t$.
In fact the output csv file has four columns (with header): first column contains the sorted values of independent parameter $t$ within the passed interval discretized by discretization step; second, third and fourth columns contain respectively the values of the components $x$, $y$ and $z$, ie the values of functions $x(t)$, $y(t)$ and $z(t)$ correspondent to values of $t$ of first column.
To get the program usage you can run this following command: 

$ python pmc3t_gen.py --help
and the output got is: 

usage: pmc3t_gen.py [-h] [--version] --dsout DS_OUTPUT_FILENAME --funcxt
                    FUNCX_T_BODY --funcyt FUNCY_T_BODY --funczt FUNCZ_T_BODY
                    [--tbegin RANGE_BEGIN] [--tend RANGE_END]
                    [--tstep RANGE_STEP] [--xnoise XNOISE_BODY]
                    [--ynoise YNOISE_BODY] [--znoise ZNOISE_BODY]

pmc3t_gen.py generates a synthetic dataset file that contains the points of a
parametric curve in space calling a a triple of one-variable real functions in
an interval

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --dsout DS_OUTPUT_FILENAME
                        dataset output file (csv format)
  --funcxt FUNCX_T_BODY
                        x=x(t) body (lamba format)
  --funcyt FUNCY_T_BODY
                        y=y(t) body (lamba format)
  --funczt FUNCZ_T_BODY
                        z=z(t) body (lamba format)
  --tbegin RANGE_BEGIN  begin range (default:-5.0)
  --tend RANGE_END      end range (default:+5.0)
  --tstep RANGE_STEP    step range (default: 0.01)
  --xnoise XNOISE_BODY  noise(sz) body (lamba format) on x
  --ynoise YNOISE_BODY  noise(sz) body (lamba format) on y
  --znoise ZNOISE_BODY  noise(sz) body (lamba format) on z
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --dsout: path (relative or absolute) of the csv file to be generated and got executing the dataset generator function.

  • --funcxt: the generator function $x=fx(t)$ of the time dataset in lamba body format where $t$ is the independent variable.

  • --funcyt: the generator function $y=fy(t)$ of the time dataset in lamba body format where $t$ is the independent variable.

  • --funcyt: the generator function $z=fz(t)$ of the time dataset in lamba body format where $t$ is the independent variable.

  • --tbegin e --tend: the closed interval of the variable $t$, between --tbegin and --tend.

  • --tstep: the discretization step of the independent variable $x$.

  • --xnoise: the noise generator functon on $x$ variable in lamba body format, where $sz$ is the number of elements of the discretized interval between --tbegin and --tend.

  • --ynoise: the noise generator functon on $y$ variable in lamba body format, where $sz$ is the number of elements of the discretized interval between --tbegin and --tend.

  • --znoise: the noise generator functon on $z$ variable in lamba body format, where $sz$ is the number of elements of the discretized interval between --tbegin and --tend.


Comparative display of datasets

This chapter explains the tools for viewing datasets allowing to visually compare the input datasets with the prediction generated by the approximators.


fx_scatter.py

Goal of the fx_scatter.py Python program is to display the scatter graph of the prediction of a real-valued scalar function of a real-valued variable superimposed on the scatter graph of the input dataset of the original function and allow the visual comparison of the two graphs.

To get the program usage you can run this following command: 

$ python fx_scatter.py --help
and the output got is: 

usage: fx_scatter.py [-h] [--version] --ds DATASET_FILENAME --prediction
                     PREDICTION_DATA_FILENAME [--title FIGURE_TITLE]
                     [--xlabel X_AXIS_LABEL] [--ylabel Y_AXIS_LABEL]
                     [--labelfontsize LABEL_FONT_SIZE] [--width WIDTH]
                     [--height HEIGHT] [--savefig SAVE_FIGURE_FILENAME]

fx_scatter.py shows two overlapped x/y scatter graphs: the blue one is the
input dataset, the red one is the prediction

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --ds DATASET_FILENAME
                        dataset file (csv format)
  --prediction PREDICTION_DATA_FILENAME
                        prediction data file (csv format)
  --title FIGURE_TITLE  if present, it set the title of chart
  --xlabel X_AXIS_LABEL
                        label of x axis
  --ylabel Y_AXIS_LABEL
                        label of y axis
  --labelfontsize LABEL_FONT_SIZE
                        label font size
  --width WIDTH         width of whole figure (in inch)
  --height HEIGHT       height of whole figure (in inch)
  --savefig SAVE_FIGURE_FILENAME
                        if present, the figure is saved on a file instead to be
                        shown on screen
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --ds: path (relative or absolute) to input dataset file; the color of its related scatter graph is blue.

  • --prediction: path (relative or absolute) to prediction file (always in csv format); the color of its related scatter graph is red.

  • --title: title of the figure, centered at the top.

  • --xlabel: $x$ axis caption.

  • --ylabel: $y$ axis caption.

  • --labelfontsize: sets the font of the texts in the figure.

  • --width: sets the width of the generated figure, in inch

  • --height: sets the height of the generated figure, in inch

  • --savefig: if it is present, it is the path (relative or absolute) of the figure file in .png format to be saved; if not present, the graph is shown on screen in a window.

An example of using the fx_scatter.py program

Having available the mytest.csv test dataset (for example generated via fx_gen.py as shown in a previous paragraph) and the mypred.csv prediction csv file (generated through a prediction program in the repository for the fitting of functions of one variable), to generate the two graphs, run the following command: 

$ python fx_scatter.py \
  --ds mytest.csv \
  --prediction mypred.csv \
  --title "My example" \
  --xlabel "x" \
  --ylabel "y=|sin(x^x)/2^(x^x-pi/2)/pi|"
that shows the two superimposed curves: in blue that of the test dataset, in red that of the prediction.

Figure with scatter graphics generated by the program fx_scatter.py showing the fitting in red overlay of the function $f(x)=|\frac{\sin x^x}{2 ^ {(x^x - \pi/2)/\pi}}|$ and the original function below in blue.

fxy_scatter.py

Goal of the fxy_scatter.py Python program is to display the scatter graph of the prediction of a real-valued scalar function of two real-valued variables alongside the scatter graph of the input dataset of the original function and allow the visual comparison of the two graphs.

To get the program usage you can run this following command: 

$ python fxy_scatter.py --help
and the output got is: 

usage: fxy_scatter.py [-h] [--version] --ds DATASET_FILENAME --prediction
                      PREDICTION_DATA_FILENAME [--title FIGURE_TITLE]
                      [--xlabel X_AXIS_LABEL] [--ylabel Y_AXIS_LABEL]
                      [--zlabel Z_AXIS_LABEL]
                      [--labelfontsize LABEL_FONT_SIZE] [--width WIDTH]
                      [--height HEIGHT] [--savefig SAVE_FIGURE_FILENAME]
fxy_scatter.py: error: the following arguments are required: --ds, --prediction
(base) malkuth:common ettore$ python fxy_scatter.py --help
usage: fxy_scatter.py [-h] [--version] --ds DATASET_FILENAME --prediction
                      PREDICTION_DATA_FILENAME [--title FIGURE_TITLE]
                      [--xlabel X_AXIS_LABEL] [--ylabel Y_AXIS_LABEL]
                      [--zlabel Z_AXIS_LABEL]
                      [--labelfontsize LABEL_FONT_SIZE] [--width WIDTH]
                      [--height HEIGHT] [--savefig SAVE_FIGURE_FILENAME]

fxy_scatter.py shows two alongside x/y/z scatter graphs: the blue one is the
surface of dataset, the red one is the surface of prediction

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --ds DATASET_FILENAME
                        dataset file (csv format)
  --prediction PREDICTION_DATA_FILENAME
                        prediction data file (csv format)
  --title FIGURE_TITLE  if present, it set the title of chart
  --xlabel X_AXIS_LABEL
                        label of x axis
  --ylabel Y_AXIS_LABEL
                        label of y axis
  --zlabel Z_AXIS_LABEL
                        label of z axis
  --labelfontsize LABEL_FONT_SIZE
                        label font size
  --width WIDTH         width of animated git (in inch)
  --height HEIGHT       height of animated git (in inch)
  --savefig SAVE_FIGURE_FILENAME
                        if present, the chart is saved on a file instead to be
                        shown on screen
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --ds: path (relative or absolute) to input dataset file; the color of its related scatter graph is blue.

  • --prediction: path (relative or absolute) to prediction file (always in csv format); the color of its related scatter graph is red.

  • --title: title of the figure, centered at the top.

  • --xlabel: $x$ axis caption.

  • --ylabel: $y$ axis caption.

  • --zlabel: $z$ axis caption.

  • --labelfontsize: sets the font of the texts in the figure.

  • --width: sets the width of the generated figure, in inch

  • --height: sets the height of the generated figure, in inch

  • --savefig: if it is present, it is the path (relative or absolute) of the figure file in .png format to be saved; if not present, the graph is shown on screen in a window.

An example of using the fxy_scatter.py program

Having available the mytest.csv test dataset (for example generated via fxy_gen.py as shown in a previous paragraph) and the mypred.csv prediction csv file (generated through a prediction program in the repository for the fitting of functions of two variables), to generate the two graphs, run the following command: 

$ python fxy_scatter.py \
  --ds mytest.csv \
  --prediction mypred.csv \
  --title "My example" \
  --xlabel "x" \
  --ylabel "y" \
  --zlabel "z=2 sin(x)^2 / sqrt(1+y^2)"
that shows the two surfaces side by side: in blue that of the input dataset, in red that of the prediction.

Figure with scatter graphics generated by the program fxy_scatter.py showing the fitting to the right in red of the function $f(x,y)=2 \frac{\sin^2 x}{\sqrt{1 + y^2}}$ and alongside the graph of the original function to the left in blue.

pmc2t_scatter.py

Goal of the pmc2t_scatter.py Python program is to display the scatter graph of the prediction of a parametric curve on plane overlapped the scatter graph of the input dataset of the original parametric curve and allow the visual comparison of the two graphs.

To get the program usage you can run this following command: 

$ python pmc2t_scatter.py --help
and the output got is: 

usage: pmc2t_scatter.py [-h] [--version] --ds DATASET_FILENAME --prediction
                        PREDICTION_DATA_FILENAME [--title FIGURE_TITLE]
                        [--xlabel X_AXIS_LABEL] [--ylabel Y_AXIS_LABEL]
                        [--labelfontsize LABEL_FONT_SIZE] [--width WIDTH]
                        [--height HEIGHT] [--savefig SAVE_FIGURE_FILENAME]

pmc2t_scatter.py shows two overlapped x/y scatter graphs on plane: the blue
one is the input dataset of parametric curve, the red one is the prediction of
parametric curve

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --ds DATASET_FILENAME
                        dataset file (csv format)
  --prediction PREDICTION_DATA_FILENAME
                        prediction data file (csv format)
  --title FIGURE_TITLE  if present, it set the title of chart
  --xlabel X_AXIS_LABEL
                        label of x axis
  --ylabel Y_AXIS_LABEL
                        label of y axis
  --labelfontsize LABEL_FONT_SIZE
                        label font size
  --width WIDTH         width of whole figure (in inch)
  --height HEIGHT       height of whole figure (in inch)
  --savefig SAVE_FIGURE_FILENAME
                        if present, the figure is saved on a file instead to
                        be shown on screen
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --ds: path (relative or absolute) to input dataset file; the color of its related scatter graph is blue.

  • --prediction: path (relative or absolute) to prediction file (always in csv format); the color of its related scatter graph is red.

  • --title: title of the figure, centered at the top.

  • --xlabel: $x$ axis caption.

  • --ylabel: $y$ axis caption.

  • --labelfontsize: sets the font of the texts in the figure.

  • --width: sets the width of the generated figure, in inch

  • --height: sets the height of the generated figure, in inch

  • --savefig: if it is present, it is the path (relative or absolute) of the figure file in .png format to be saved; if not present, the graph is shown on screen in a window.


pmc3t_scatter.py

Goal of the fxy_scatter.py Python program is to display the scatter graph of the prediction of a parametric curve in space overlapped the scatter graph of the input dataset of the original parametric curve and allow the visual comparison of the two graphs.

To get the program usage you can run this following command: 

$ python pmc3t_scatter.py --help
and the output got is: 

  -h, --help            show this help message and exit
  --version             show program&aposMs version number and exit
  --ds DATASET_FILENAME
                        dataset file (csv format)
  --prediction PREDICTION_DATA_FILENAME
                        prediction data file (csv format)
  --title FIGURE_TITLE  if present, it set the title of chart
  --xlabel X_AXIS_LABEL
                        label of x axis
  --ylabel Y_AXIS_LABEL
                        label of y axis
  --zlabel Z_AXIS_LABEL
                        label of z axis
  --labelfontsize LABEL_FONT_SIZE
                        label font size
  --width WIDTH         width of whole figure (in inch)
  --height HEIGHT       height of whole figure (in inch)
  --savefig SAVE_FIGURE_FILENAME
                        if present, the figure is saved on a file instead to
                        be shown on screen
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --ds: path (relative or absolute) to input dataset file; the color of its related scatter graph is blue.

  • --prediction: path (relative or absolute) to prediction file (always in csv format); the color of its related scatter graph is red.

  • --title: title of the figure, centered at the top.

  • --xlabel: $x$ axis caption.

  • --ylabel: $y$ axis caption.

  • --zlabel: $z$ axis caption.

  • --labelfontsize: sets the font of the texts in the figure.

  • --width: sets the width of the generated figure, in inch

  • --height: sets the height of the generated figure, in inch

  • --savefig: if it is present, it is the path (relative or absolute) of the figure file in .png format to be saved; if not present, the graph is shown on screen in a window.


Diagnostics

This chapter explains the diagnostic tools provided by this repository for the fitting of functions; the programs in this repository that implement the training features save (if the user so wishes) dump files usable later and/or during the training phase by the programs described herein with the aim of analyzing how the training phase has happened (or is happening) and then operate on the hyper parameters to improve the final result.


dumps_scatter.py

Goal of the dumps_scatter.py Phyton program is to display the graphs of the loss function and the various metrics calculated during the training phase. In order to be able to use this program it is necessary that the training programs have been launched with the appropriate arguments to save the dump files of the loss function and/or metrics as the epochs change.

To get the program usage you can run this following command: 

$ python dumps_scatter.py --help
and the output got is: 

usage: dumps_scatter.py [-h] [--version] --dump DUMP_PATH
                        [--savefigdir SAVE_FIGURE_DIRECTORY]

dumps_scatter.py shows the loss and metric graphs with data generated by
any fitting program with argument --dumpout

optional arguments:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --dump DUMP_PATH      Dump directory (generated by any fitting/training
                        programs of this suite that support --dumpout
                        argument)
  --savefigdir SAVE_FIGURE_DIRECTORY
                        If present, the charts are saved on files in
                        savefig_dir folder instead to be shown on screen
Where:
  • -h, --help: shows the usage of the program and ends the execution.

  • --version: shows the version of the program and ends the execution.

  • --dump: path (relative or absolute) of the directory where the loss function and/or metrics dumps were saved during the training as the epochs change.

  • --savefigdir: if it is present, it is the path (relative or absolute) of the directory where the program saves the figures in .png format, one figure for each graph for which the corresponding dump file exists; if not present, the graphs are shown on screen.

An example of using the dumps_scatter.py program

Having available the directory mydumps which contains the dump files of some metrics calculated during the training phase of some training program in this repository, to view the graphs of these metrics run the following command: 

$ python dumps_scatter.py \
  --dump mydumps
which shows in sequence the graphs of the metrics saved in the mydumps folder, as shown in the following figures.

Graph of the metric logloss as the epochs change calculated on the training dataset.

Graph of the metric error as the epochs change calculated on the training dataset.

Graph of the metric logloss as the epochs change calculated on validation dataset.

Graph of the metric error as the epochs change calculated on validation dataset.


Download of the complete code

The complete code is available at GitHub.
These materials are distributed under MIT license; feel free to use, share, fork and adapt these materials as you see fit.
Also please feel free to submit pull-requests and bug-reports to this GitHub repository or contact me on my social media channels available on the top right corner of this page.