Skip to main content

CoTorch

A machine learning framework built from first principles.

CoTorch is an experimental, open-source machine learning framework focused on understanding and implementing the foundations of modern ML.

The project explores the full stack, from tensor operations and automatic differentiation to neural networks, optimisation algorithms, and model training.

Rather than competing with established frameworks, CoTorch serves as a space for hands-on implementation, experimentation, and exploration of machine learning systems.

Areas of exploration

  • Tensor computation: Multidimensional arrays, broadcasting, memory layouts, and numerical operations.
  • Automatic differentiation: Computational graphs, reverse-mode autodiff, and gradient computation.
  • Neural networks: Layers, activation functions, loss functions, and optimisation algorithms.
  • Machine learning algorithms: Implementations of classical and modern ML methods.
  • Performance: CPU/GPU computation, numerical efficiency, and benchmarking.

Development

CoTorch is in its early stages of development.

The initial focus is on building the core tensor and numerical computation primitives, followed by automatic differentiation and neural network training.

The project is exploratory, and its scope will evolve as new components are implemented.

  • CoInfer: An experimental ML inference engine exploring model execution and optimisation.
  • CoHarness: An experimental AI coding-agent harness exploring model integration, tool calling, context management, and autonomous execution.

License

MIT License.

Metadata

Release files for cotorch 0.0.1

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for cotorch 0.0.1
File Size Uploaded Provenance Actions
cotorch-0.0.1.tar.gz 11.4 kB Without provenance Details for cotorch-0.0.1.tar.gz

Release files / cotorch-0.0.1.tar.gz

Download URL cotorch-0.0.1.tar.gz
Size 11.4 kB
Tags Source
SHA-256 checksum
How to use checksums
7bf0a219defc453cb997e68762e82ccceeb365e6781508981070b2046b27f2e4
BLAKE2b-256 checksum
How to use checksums
0fd5473cac829d690a39f3bbae720a1b2b7a952bfae06ccf8a97799cbd2bb059
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/7.0.0 CPython/3.14.7

Security for cotorch 0.0.1

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Release history Release notifications | RSS feed

This release

0.0.1 This release

1 release file No provenance

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page