Welcome to my Machine Learning repository!
This repository contains my practice implementations of core AI/ML concepts built completely from scratch to better understand the mathematics, gradients, backpropagation, and internal workings behind modern machine learning systems.
The goal of this repository is to gain a deeper understanding of how machine learning models work internally by implementing algorithms manually rather than relying entirely on high-level abstractions and frameworks.
Verified machine learning implementations completed on TensorTonic.
| Problem | Description | Link |
|---|---|---|
| Compute Entropy for a Node | Compute decision-tree node entropy from class labels using empirical class probabilities and base-two logarithms. | https://www.tensortonic.com/problems/entropy-node |
| Logistic Regression Training Loop | Train binary logistic regression in NumPy using sigmoid probabilities, gradient descent, and learned weight and bias parameters. | https://www.tensortonic.com/problems/logistic-regression-training |
| Matrix Transpose | Implement matrix transpose in NumPy without built-in transpose helpers, preserving rectangular shapes and the original input. | https://www.tensortonic.com/problems/matrix-transpose |
| Pad Sequences | Pad or truncate variable-length token ID sequences in NumPy with configurable maximum length and padding values. | https://www.tensortonic.com/problems/pad-sequences |
| RMSProp Optimizer (Single Update Step) | Implement one RMSProp update in NumPy using an exponential squared-gradient average and adaptive scaling. | https://www.tensortonic.com/problems/rmsprop-optimizer |
| Implement Sigmoid in NumPy | Implement a vectorized sigmoid activation in NumPy for scalars, lists, vectors, and matrices, including large positive and negative inputs. | https://www.tensortonic.com/problems/sigmoid-numpy |
| Implement Softmax Function | Implement numerically stable softmax by shifting logits before exponentiation and normalizing probabilities. | https://www.tensortonic.com/problems/softmax-function |
| Leaky ReLU | Implement Leaky ReLU activation in CUDA with one thread per element, bounds checks, and a configurable negative slope. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/leaky-relu |
| Matrix Addition | Implement elementwise matrix addition in CUDA with a two-dimensional grid, row-major indexing, and bounds checks. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/matrix-addition |
| Matrix Transpose | Implement matrix transpose in CUDA with a two-dimensional launch grid, row-major buffers, and bounds-checked writes. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/matrix-transpose |
| ReLU | Implement ReLU activation in CUDA with one thread per element, bounds checks, and branch-efficient rectification. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/relu |
| Sigmoid | Implement sigmoid activation in CUDA with one thread per element, device exponential math, and bounds-checked memory access. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/sigmoid |
| Sum of Array | Implement a multi-block CUDA sum reduction that combines partial block sums into one scalar output. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/sum-of-array |
| Tanh | Implement hyperbolic tangent activation in CUDA with one thread per element, device intrinsic math, and bounds checks. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/tanh |
| Vector Addition | Implement bounds-checked pointwise vector addition in CUDA with one thread per output element. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/vector-addition |
| Vector Subtraction | Implement bounds-checked pointwise vector subtraction in CUDA with one thread per output element. | https://www.tensortonic.com/study-plans/cuda-basics/cuda/vector-subtract |
View my verified ML profile: TensorTonic profile