ocelotl-kernels 0.0.1

Portable kernel dispatch boundary for Ocelotl.
Documentation
  • Coverage
  • 12.5%
    1 out of 8 items documented0 out of 6 items with examples
  • Size
  • Source code size: 5.83 kB This is the summed size of all the files inside the crates.io package for this release.
  • Documentation size: 385.51 kB This is the summed size of all files generated by rustdoc for all configured targets
  • Ø build duration
  • this release: 22s Average build duration of successful builds.
  • all releases: 22s Average build duration of successful builds in releases after 2024-10-23.
  • Links
  • Dieshen/ocelotl
    0 0 0
  • crates.io
  • Dependencies
  • Versions
  • Owners
  • Dieshen

Ocelotl

Rust-first LLM inference runtime.

Ocelotl is an early-stage workspace for a local LLM runtime with explicit model, loader, kernel, and serving boundaries. The first milestone is a narrow, correct single-process runtime before adding broad model coverage or high-scale serving features.

Crates

  • ocelotl-core: shared types, errors, model metadata, and device contracts.
  • ocelotl-loader: model artifact loading and validation.
  • ocelotl-tokenizer: tokenizer and chat-template boundary.
  • ocelotl-kernels: portable kernel dispatch boundary.
  • ocelotl-models: model-family implementations.
  • ocelotl-runtime: request lifecycle, KV cache, scheduling, and generation.
  • ocelotl-server: API/server integration layer.
  • ocelotl: root crate and CLI entrypoint.

Current Status

This is a project skeleton. Public APIs are intentionally small while the runtime shape is established.