mini-torch

Docs : https://sirohikartik.github.io/mini-torch/docs.html

mini-torch is a minimal deep learning framework built from scratch in Python. It implements reverse-mode automatic differentiation, a dynamic computation graph, and a basic neural network training loop.

The project is designed for learning and experimentation, not performance.

---

Features

• Custom Tensor class with:
  • Dynamic computation graph
  • Reverse-mode automatic differentiation
  • Gradient accumulation
• Supported operations:
  • Addition and subtraction
  • Matrix multiplication
  • ReLU activation
  • Sum reduction
• Backpropagation via topological sorting
• SGD optimizer with momentum
• Adam optimizer
• Simple multi-layer perceptron (MLP)
• End-to-end training loop

---

Example

from tensor import Tensor
import numpy as np

a = Tensor([[1.0, 2.0]])
b = Tensor([[3.0], [4.0]])

c = a @ b
loss = c.sum()
loss.backward()

print(a.grad)
print(b.grad)