ttglow 0.1.1

Tensor Train linear algebra package for efficient high-dimensional tensor operations

TTGlow

A Tensor Train linear algebra package for efficient high-dimensional tensor operations in PyTorch.

Overview

TTGlow provides an efficient implementation of the Tensor Train (TT) decomposition, a powerful technique for representing and manipulating high-dimensional tensors with reduced memory footprint and computational cost. This package is particularly useful for:

• High-dimensional data analysis
• Quantum many-body physics simulations
• Machine learning with tensor methods
• Efficient storage and computation of large-scale tensors

Installation

Using Pixi (Recommended)

# Clone the repository
git clone https://github.com/romanellerbrock/ttglow.git
cd ttglow

# Install dependencies with pixi
pixi install

# Activate the environment
pixi shell

Using pip

pip install -e .

Quick Start

import torch
from ttglow import TensorTrain, dot, add, scale

# Create a random Tensor Train
dims = [10, 20, 15]  # Tensor dimensions
ranks = [3, 5]        # TT-ranks
tt = TensorTrain.random(dims, ranks)

# Convert to full tensor
full_tensor = tt.to_tensor()

# Create a constant tensor in TT format
tt_ones = TensorTrain.ones([10, 20, 15], value=3.14)

# Dot product (inner product)
tt1 = TensorTrain.random([4, 5, 6], [2, 3])
tt2 = TensorTrain.random([4, 5, 6], [2, 3])
matrices = dot(tt1, tt2)
inner_product = matrices[-1].item()  # Scalar result
norm_squared = dot(tt1, tt1)[-1].item()  # Self inner product

# Addition
tt_sum = add(tt1, tt2)  # Returns new TT with ranks = r1 + r2

# Scalar multiplication
tt_scaled = scale(tt1, 2.5)  # Multiply by scalar

Features

• Efficient TT representation: Store high-dimensional tensors with minimal memory
• Core operations:
  • dot: Inner product of two TT tensors
  • hadamard: Element-wise multiplication
  • add: Summation of TT tensors
  • scale: Scalar multiplication
• Flexible construction: Create TT tensors from scratch or via decomposition
• PyTorch backend: Full GPU support and automatic differentiation

Development

Running Tests

pixi run test

Code Formatting

pixi run format

Linting

pixi run lint

Test Coverage

pixi run test-cov

Project Structure

ttglow/
├── src/ttglow/           # Main package code
│   ├── __init__.py
│   └── tensortrain.py    # TT implementation
├── tests/                # Test suite
│   └── test_tensortrain.py
├── examples/             # Usage examples
├── pixi.toml             # Pixi configuration
└── pyproject.toml        # Python package metadata

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

References

• Oseledets, I. V. (2011). "Tensor-train decomposition". SIAM J. Sci. Comput., 33(5), 2295-2317.