Headline
- FastFlowLM's NPU models and runtime have been integrated. Switch between ONNX, GGUF, and FLM models with 1 click.
- macOS and Apple silicon are supported via llamacpp+Metal.
- Lots of awesome community contributions that overhauled the web ui and more.
What's Changed
- Add model size metadata (GB) to server_models and UI Model Manager by @siavashhub in #386
- Add vision checkbox to model registration form by @Deepam02 in #383
- Initial macOS support for lemonade server by @ramkrishna2910 in #290
- Update docs with macos support by @ramkrishna2910 in #391
- Fix website by @ramkrishna2910 in #395
- Fix 334 installer text by @ajnatopic1 in #397
- Fix major CI blocking issues by @danielholanda in #418
- Return a 422 when responses API is sent to llamacpp server by @jeremyfowers in #402
- Optimize macos tests by @ramkrishna2910 in #412
- Feat/UI stop button by @kpoineal in #370
- Update Llama.cpp+ROCm to use ROCWMMA by @danielholanda in #396
- Update hot models list by @jeremyfowers in #427
- Use an image for the lemonade logo on the web ui by @jeremyfowers in #428
- Fix text paste bug by @jeremyfowers in #429
- Fix: WebUI not expanding as expected by @siavashhub in #400
- Move the model status indicator inside the dropdown select ui element by @jeremyfowers in #426
- Skip longer tests for macos to reduce CI time by @ramkrishna2910 in #433
- Integrate FastFlowLM (FLM) by @jeremyfowers in #409
- Web UI Server Log Feature (#179) by @siavashhub in #431
- Fix model manager button status during long tasks by @jeremyfowers in #424
New Contributors
- @siavashhub made their first contribution in #386
- @Deepam02 made their first contribution in #383
- @ajnatopic1 made their first contribution in #397
Full Changelog: v8.1.10...v8.1.11