Getting Started¶

This guide walks through installation, backend setup, and your first steps with tinyTT.

Installation¶

Prerequisites¶

Python 3.11+
tinygrad (default backend) or PyTorch (optional)

Install from Source¶

git clone https://github.com/meigel/tinyTT.git
cd tinyTT
python3 -m venv venv && source venv/bin/activate
pip install -r requirements.txt
pip install -e .

Optional: GPU Support via Pinned tinygrad¶

For GPU acceleration (CUDA, Metal, OpenCL), install the pinned tinygrad submodule rather than the PyPI wheel:

git submodule update --init tinygrad
pip install ./tinygrad

Optional: PyTorch Backend¶

pip install torch
TINYTT_BACKEND=pytorch python my_script.py

Development Dependencies¶

pip install -r requirements-dev.txt    # pytest

Your First TT Tensor¶

import numpy as np
import tinytt as tt
import tinytt._backend as tn

# Build a TT tensor from a full 2×2×2 array
full = np.arange(8, dtype=np.float64).reshape(2, 2, 2)
x = tt.TT(full, eps=1e-12)
print("TT ranks:", x.R)                    # [1, 1]
print("Shape:", x.N)                       # [2, 2, 2]

# Materialise back to dense
recon = tn.to_numpy(x.full())
rel_err = np.linalg.norm(recon - full) / np.linalg.norm(full)
print(f"Reconstruction error: {rel_err:.3e}")   # ≈ 0

TT-Matrix and Matvec¶

# Identity TT-matrix
A = tt.eye([4, 4])
x = tt.ones([4, 4])
b = A @ x
print((b - x).norm().numpy())    # ≈ 0

Switching Backends¶

Set the TINYTT_BACKEND environment variable:

TINYTT_BACKEND=pytorch python my_script.py
TINYTT_BACKEND=tinygrad python my_script.py     # default

All code is backend-agnostic — import the facade via import tinytt._backend as tn rather than calling tinygrad or PyTorch directly.

Next Steps¶

Follow the TT Basics tutorial for a deeper dive into construction, rounding, and decomposition.
Browse Examples for ready-to-run scripts.
Check the source code on GitHub for detailed module documentation.