Make a unit test for the MLP class

[benlyazid]

Title: Build Unit Test Cases for the MLP Class

Description:

Create a comprehensive set of unit test cases to validate the functionality of the MLP class. The tests should cover different architectures, scenarios, and data types to ensure robustness and correctness.

Tasks:

1. Test Different Architectures:

   • Single-layer vs. multi-layer architectures
   • Different numbers of neurons per layer
   • Different activation functions (ReLU, Sigmoid, Tanh)

2. Test Different Scenarios:

   • Small vs. large datasets
   • Edge cases (e.g., very small/large inputs, all zeros, negative values)
   • Handling of batch sizes (e.g., batch size of 1, batch size larger than dataset)

3. Test with Different Data Types:

   • Integer, float, and mixed-type inputs
   • Normalized vs. non-normalized data

4. Validate Expected Behavior:

   • Ensure forward propagation outputs match expected values
   • Check loss reduction over epochs in training
   • Validate gradients in backpropagation