WIP - A simple C library to work with machine learning

---

The goal is to provide header-only library that you can use that anywhere.

Note: RAII is implemented using __attribute__((cleanup)), which only works in Clang and GCC. Manual memory management is still available through FreeArray(array), FreeArrayPointer(array_pointer), and Array(T, size).

Note2: Both src/main.c and src/bench.py was done using AI, it's just to benchmark the lib and compare to numpy, i still need a better example.

Todo(s)

• Update the functions models to use destination buffers, e.g., func(src1, src2, dest). This enforces stricter memory management and reduces leaks.
• Add support for hardware offloading (GPU) or handling very large arrays in chuncks.
• Add documentation/comments for all functions (PARTIAL DONE)
• Implement a simple neural network to benchmark performance. (PARTIAL DONE)
• Try changing from RAII to a Arena allocator or support both methods.
• Remove .data field in Matrix and maintain only allocator.