Yule

Yule is a software tool designed to help people run artificial intelligence models on their own computers. It is built using the Rust programming language and focuses on security and trust. Yule allows users to get verifiable proof that the AI model they are running is exactly as intended and hasn't been tampered with.

Benefits

Yule offers several advantages. It provides cryptographic integrity, meaning you can trust that the AI model is running correctly. It allows for local AI model execution, giving you more control. The software is designed with security in mind, using features like kernel-level sandboxing to protect your system. It also offers GPU acceleration for faster performance, up to 8.5 times faster than CPU on certain models, without being limited to specific hardware like NVIDIA. Yule also creates signed records of each AI task, making them tamper-evident.

Use Cases

Yule can be used to run various AI models locally, such as those for text generation or other tasks. Users can install models from platforms like HuggingFace and then run them for inference. It can also serve as an API server, offering both Yule's secure endpoints and standard OpenAI-compatible endpoints for easier integration with other tools. The tool also includes features for verifying model integrity and auditing past inference sessions.

Pricing

Information about pricing is not available in the provided text.

Vibes

Information about public reception or reviews is not available in the provided text.

Additional Information

Yule is currently in beta after about 2.5 months of development by a single person. It has been tested with the TinyLlama 1.1B model on both CPU and Vulkan. The project is actively working on supporting multiple AI models. Yule is licensed under the Apache-2.0 license.