두 개의 tex 파일을 병합하여 각각 고유한 참조를 갖는 방법

두 개의 tex 파일을 병합하여 각각 고유한 참조를 갖는 방법

내 논문에 보충 파일이 있는데 arXiv에 제출하고 싶습니다. 먼저 Adobe 소프트웨어를 사용하여 PDF로 변환하고 병합해 보았습니다. 그러나 arXiv는 실제 소스 코드를 보내도록 강요합니다. 문제로 돌아가서 간단히 사용하면

...
\bibliographystyle{ieeetr}
{ \scriptsize
\bibliography{refs} 
}

\newpage
\input{SuppMaterial} 

내가 원하는 것을 얻지 못했습니다. 내 SuppMaterial에도 참조가 있으므로 refs 블록이 반복된다는 뜻입니다. SuppMaterial의 끝에서도 전체 참조를 얻습니다.
각 파일이 자체 참조 블록을 갖도록 이 두 개를 어떻게 병합할 수 있습니까?(첫 번째 파일은 해당 참조로 시작하고 끝나고, 두 번째 파일은 자체 참조로 시작하고 끝납니다.)

더 정확하게 말하면, 이 두 파일의 모양에 대한 예는 main.tex다음과 같습니다.

\documentclass[journal]{IEEEtran}
\usepackage{subcaption}     % for subfigures
\usepackage[utf8]{inputenc} % allow utf-8 input
\usepackage[T1]{fontenc}    % use 8-bit T1 fonts
\usepackage{hyperref}       % hyperlinks
\usepackage{url}            % simple URL typesetting
\usepackage{booktabs}       % professional-quality tables
\usepackage{amsfonts}       % blackboard math symbols
\usepackage{nicefrac}       % compact symbols for 1/2, etc.
\usepackage{microtype}      % microtypography
\usepackage{epsfig}
\usepackage{graphicx}
\hypersetup{colorlinks=true, linkcolor=red}

\usepackage{algorithm}
\usepackage{algorithmic}
\usepackage{tikz}
\usepackage{pgfplots}
\usepackage{amsmath}
\usepackage{sci}
\usepackage{xfrac}
\usepackage{pgfplots}
\usepackage{pgfplotstable}
\usepackage{filecontents,pgfplots}
\usepackage{tikz}
\usetikzlibrary{arrows,calc,shapes,snakes,positioning,matrix,arrows,decorations.pathmorphing,decorations.text}
\usepackage[sort&compress,square,comma,numbers]{natbib}
\makeatletter
\renewcommand\bibsection%
{
  \section*{\refname
    \@mkboth{\MakeUppercase{\refname}}{\MakeUppercase{\refname}}}
}
\makeatother

% correct bad hyphenation here
\hyphenation{op-tical net-works semi-conduc-tor}


\begin{document}

%
% paper title

\title{Title}
%

%\author{XXX, YYY,ZZZ%,~\IEEEmembership{Member,~IEEE,}% <-this % stops a space
}


% use for special paper notices
%\IEEEspecialpapernotice{(Invited Paper)}




% make the title area
\maketitle


\begin{abstract}
Here is the abstract
\end{abstract}

% Note that keywords are not normally used for peerreview papers.
\begin{IEEEkeywords}
Deep Learning, Convolutional Neural Networks (CNNS), Efficiency. 
\end{IEEEkeywords}

\section{Introduction}

...

\bibliographystyle{ieeetr}
{ \scriptsize
\bibliography{refs} 
}
\end{document}

supp.tex다음과 같이 보입니다.

\documentclass[journal]{IEEEtran}

\usepackage{subcaption}     % for subfigures
\usepackage[utf8]{inputenc} % allow utf-8 input
\usepackage[T1]{fontenc}    % use 8-bit T1 fonts
\usepackage{hyperref}       % hyperlinks
\usepackage{url}            % simple URL typesetting
\usepackage{booktabs}       % professional-quality tables
\usepackage{amsfonts}       % blackboard math symbols
\usepackage{nicefrac}       % compact symbols for 1/2, etc.
\usepackage{microtype}      % microtypography
\usepackage{epsfig}
\usepackage{graphicx}
\hypersetup{colorlinks=true, linkcolor=red}

\usepackage{tabulary}

\usepackage{algorithm}
\usepackage{algorithmic}
\usepackage{tikz}
\usepackage{pgfplots}
\usepackage{amsmath}
\usepackage{sci}
\usepackage{xfrac}
\usepackage{pgfplots}
\usepackage{pgfplotstable}
\usepackage{filecontents,pgfplots}
\usepackage{tikz}
\usetikzlibrary{arrows,calc,shapes,snakes,positioning,matrix,arrows,decorations.pathmorphing,decorations.text}

% To number supplemental material with 'S': 
\renewcommand{\thepage}{S\arabic{page}}  
\renewcommand{\thesection}{S\arabic{section}}   
\renewcommand{\thetable}{S\arabic{table}}   
\renewcommand{\thefigure}{S\arabic{figure}}


\begin{document}

\title{Title}

%\author{X, Y, Z,%,~\IEEEmembership{Member,~IEEE,}% <-this % stops a space


\maketitle

% \begin{abstract}
% \end{abstract}


\section{Discussions}

\subsection{Subject 1}

...

\ifCLASSOPTIONcaptionsoff
  \newpage
\fi


\bibliographystyle{ieeetr}
{ \scriptsize
\bibliography{refs} 
}
% \begin{thebibliography}{10}

% \end{thebibliography}

\end{document}

두 가지 모두에서 사용되는 ref.bib는 다음과 같습니다.

@inproceedings{AlexKrizhevsky_imgnet_2012,
   author = {Alex, Krizhevsky and Sutskever, Ilya and Geoffrey, E. Hinton},
   title = {ImageNet Classification with Deep Convolutional Neural Networks},
   pages = {1097--1105},
   booktitle = {NIPS},
   year = {2012},
}

@inproceedings{Ba_DoDeepNets_nd_B_Deep_2014,
   author = {Ba, Jimmy and Caruana, Rich},
   title = {Do deep nets really need to be deep?},
   booktitle = {NIPS},
   pages = {2654-2662},
   type = {Conference Proceedings}
}

@inproceedings{Bucilua_Model_compression_2006,
   author = {Buciluǎ, Cristian and Caruana, Rich and Niculescu-Mizil, Alexandru},
   title = {Model compression},
   booktitle = {SIGKDD},
   publisher = {ACM},
   pages = {535-541},
   ISBN = {1595933395},
   type = {Conference Proceedings}
}

관련 정보