Issue combining makecell and cellcolor

The problem with makecell and colouring tables has been known for quite some time. A simple possibility, consists in replacing \makecell in coloured columns/rows with \Centerstack from stackengine. I took the opportunity to simplify a bit your code (the default input encoding is now utf8, for instance, so needless to load inputenc) and added some improvelments, Stype columns for the alignment on the decimal dot, and loading of v=cellspace to add some vertical padding to multiline cells in the first column.

   \documentclass[12pt, a4paper,article]{memoir}
    \usepackage[T1]{fontenc}
    \usepackage[french,english]{babel}
    \usepackage[margin=2cm]{geometry}
    \usepackage{caption}
    \usepackage{siunitx} 
    \usepackage[dvipsnames, table, svgnames]{xcolor}
    \usepackage{diagbox}
    \usepackage{makecell}
    \usepackage[usestackEOL]{stackengine}
    \usepackage[column=O]{cellspace}
    \setlength{\cellspacetoplimit}{2pt}
    \setlength{\cellspacebottomlimit}{2pt}
    \usepackage{adjustbox}
    \newcommand\xrowht[2][0]{\addstackgap[.5\dimexpr#2\relax]{\vphantom{#1}}}

    \begin{document}

    \begin{table}[ht]
    \centering
    \sisetup{table-format=1.2, table-column-width = 4em, table-number-alignment=center}
    \begin{tabular}{| >{\columncolor{LightGrey}}Oc|*{5}{S|}}\hline
    \rowcolor{LightGrey}
   \multicolumn{1}{|c|}{ \backslashbox[50mm]{Algorithms}{Metrics}}
    &{Accuracy} &{Precision} & {Recall} &{F1 score} &{F2 score}\\
    \hline
    \cellcolor{LightGrey}\Centerstack{Neural Network \\ with Embedding} & 0.62 & 0.23 & 0.82 & 0.36 & 0.54\\
    \hline
    \Centerstack{Neural Network \\ without Embedding} & 0.64 & 0.22 & 0.69 & 0.34 & 0.49\\
    \hline
    \Centerstack{Support Vector Machine \\ (SVM)} & 0.64 & 0.21 & 0.65 &0.32 & 0.46\\
    \hline\xrowht{20pt}
    Logistic Regression & 0.60 & 0.17 & 0.55 & 0.26 & 0.38\\
    \hline\xrowht{20pt}
    Decision tree & 0.72 & 0.25 & 0.56 & 0.34 & 0.46\\
    \hline\xrowht{20pt}
    Random Forest & 0.77 & 0.27 & 0.43 & 0.33 & 0.38\\
    \hline\xrowht{20pt}
     XGBoost & 0.71 & 0.27 & 0.75 & 0.4 & 0.56\\
    \hline\xrowht{20pt}
     CatBoost & 0.70 & 0.27 & 0.75 & 0.39 & 0.55\\
    \hline
    \end{tabular}
    \caption{Results using Weight Attention Data Set.}
    \label{tab:tableweightattention}
    \end{table}

    \end{document} 

With {NiceTabular} of nicematrix:

\documentclass{article}
\usepackage[top=2cm, bottom=2cm, left=2cm, right=2cm]{geometry}
\usepackage[svgnames]{xcolor}
\usepackage{nicematrix}

\begin{document}

\NiceMatrixOptions{cell-space-limits=1mm}

\begin{table}[ht]
\centering
\begin{NiceTabular}{c*{5}{wc{4em}}}%
  [ code-before = \columncolor{LightGrey}{1} \rowcolor{LightGrey}{1} , hvlines ]
\diagbox{Algorithms}{Metrics}
&Accuracy &Precision &Recall &F1 score &F2 score \\
\Block{}{Neural Network \\ with Embedding} & 0.62 & 0.23 & 0.82 & 0.36 & 0.54\\
\Block{}{Neural Network \\ without Embedding} & 0.64 & 0.22 & 0.69 & 0.34 & 0.49\\
\Block{}{Support Vector Machine \\ (SVM)} & 0.64 & 0.21 & 0.65 &0.32 & 0.46\\
Logistic Regression & 0.60 & 0.17 & 0.55 & 0.26 & 0.38\\
Decision tree & 0.72 & 0.25 & 0.56 & 0.34 & 0.46\\
Random Forest & 0.77 & 0.27 & 0.43 & 0.33 & 0.38\\
XGBoost & 0.71 & 0.27 & 0.75 & 0.4 & 0.56\\
CatBoost & 0.70 & 0.27 & 0.75 & 0.39 & 0.55\\
\end{NiceTabular}
\caption{Results using Weight Attention Data Set.}
\end{table}

\end{document}