index.html

<!doctype html>
<html lang="en">
  <head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1" />
    <meta
      name="description"
      content="TensorFold provides fast local LLM inference and tested model releases for Apple Silicon and NVIDIA CUDA."
    />
    <title>TensorFold | Local LLM inference</title>
  </head>
  <body>
    <main>
      <header>
        <p><strong>OPEN-SOURCE LOCAL INFERENCE</strong></p>
        <img
          src="https://huggingface.co/spaces/TensorFold/README/resolve/main/tensorfold-logo.png"
          alt="TensorFold"
          width="120"
        />
        <h1>TensorFold</h1>
        <p><strong>Fast local LLM inference and tested model releases for Apple Silicon and NVIDIA CUDA.</strong></p>
        <p>
          <a href="https://tensorfold.dev">tensorfold.dev</a>
          ·
          <a href="https://github.com/ashhart/TensorFold">GitHub</a>
        </p>
      </header>

      <br />

      <p>
        TensorFold is an open-source inference runtime with an OpenAI-compatible API. This
        Hugging Face organization publishes compatible checkpoints and supporting assets tested
        on real hardware.
      </p>

      <br />

      <h2>What you will find here</h2>
      <ul>
        <li><strong>Apple Silicon:</strong> MLX quantized checkpoints with measured speed and memory use.</li>
        <li><strong>Drafted decoding:</strong> MTP and DFlash assets when the upstream model provides a compatible drafter.</li>
        <li><strong>NVIDIA CUDA:</strong> tested recipes for supported GPUs and DGX Spark systems.</li>
        <li><strong>Clear release notes:</strong> runtime versions, setup steps, known limits, licences, and upstream credit.</li>
      </ul>

      <br />

      <h2>Run TensorFold</h2>
      <p><code>curl -fsSL https://tensorfold.dev/install.sh | sh</code></p>

      <br />

      <p>
        TensorFold does not train the base models. Model design, training, evaluations, and
        upstream documentation remain the work of the original authors and contributors.
      </p>
    </main>
  </body>
</html>