I need a complete Jupyter Notebook implementing, training, and evaluating neural network models for image classification using the CIFAR-10 dataset in PyTorch. The task includes: Downloading and preparing the CIFAR-10 dataset (train, validation, and test sets). Implementing two CNN models (m2 and m3) with different convolutional layers. Training and validating both models until reaching at least 75% accuracy. Displaying results: loss and accuracy curves, confusion matrices, and evaluation metrics. Comparing model performances (m2 vs. m3). Improving the weaker model using one or more of the following: Normalization / feature scaling WeightedRandomSampler for class imbalance Dropout and pooling layers Writing clear Markdown explanations and code comments throughout the notebook. All functions (e.g., ModelBuilder, weighted_sampler) should be placed in a separate folder called handy_code. Deliverables Jupyter Notebook (.ipynb) with all working code, results, and analysis. handy_code folder containing helper functions and classes. The notebook must run without errors and use a fixed random seed for reproducibility Acceptance criteria • Notebook executes top-to-bottom on first run and reaches ≥70 % test accuracy on CIFAR-10 • Clear section headings: Data Prep, Model, Training, Evaluation, Conclusions • Plots for training/validation accuracy and loss saved to disk and displayed inline That’s the entire scope—once those points are met, the job is done.