Please use google colab or python to code this, thank you!!! I tried using AI $($chat gpt or blackbox$),$ but someone who knows machine learning, please help!!!!

MNIST classification using SVM

$1.$Select the first $60000$ samples for training, and the next $10000$ for testing.

$2.$Use linear SVM to classify MNIST digits with C values $=10^(-5$:$0\mathrm{.}5$:$5).$ Determine both training and testing accuracy values for each model and plot them versus C values with a logarithm scale.

$3.$Use PCA and linear SVM to classify MNIST digits with a C value corresponding to the best performance determined in step $2.$ The number of principal components should be equal to $50$:$50$:$700.$ Determine both training and testing accuracy values for each model and plot them versus the number of principal components.

$4.$Use cubic$-$polynomial SVM to classify MNIST digits with C values $=10^(-5$:$0\mathrm{.}5$:$5).$ Determine both training and testing accuracy values for each model and plot them versus C values with a logarithm scale.

$5.$MNIST classification using SVM

Select the first $60000$ samples for training, and the next $10000$ for testing.

Use linear SVM to classify MNIST digits with C values $={10}^{-5\text{:}0\mathrm{.}5\text{:}5}.$ Determine both training and testing

accuracy values for each model and plot them versus $C$ values with a logarithm scale.

Use PCA and linear SVM to classify MNIST digits with a C value corresponding to the best performance

determined in step $2.$ The number of principal components should be equal to $50$:$50$:$700.$ Determine both

training and testing accuracy values for each model and plot them versus the number of principal

components.

Use cubic$-$polynomial SVM to classify MNIST digits with C values $={10}^{-5\text{:}0\mathrm{.}5\text{:}5}.$ Determine both training

and testing accuracy values for each model and plot them versus $C$ values with a logarithm scale.

Use PCA and cubic$-$polynomial SVM to classify MNIST digits with a C value corresponding to the best

performance determined in step $4.$ The number of principal components should be equal to $50$:$50$:$700.$

Determine both training and testing accuracy values for each model and plot them versus the number of

principal components.Use PCA and cubic$-$polynomial SVM to classify MNIST digits with a C value corresponding to the best performance determined in step $4.$ The number of principal components should be equal to $50$:$50$:$700.$ Determine both training and testing accuracy values for each model and plot them versus the number of principal components.