🔥 What's New: v0.19.0
- 🎯 Accelerated Logit Masking: Enables faster constrained decoding on GPUs, with full support for MTP-enabled decoding.
- 🌐 Web SDK: @litert-lm/core now stream-loads all models, defaults to the GPU backend, and supports image and audio inputs.
- 🎙️ Speech Models: Added support for multiple ASR and TTS models, including Qwen3-ASR, Kokoro, and Qwen3-TTS.
- 🧱 C API 1.0.0 (Breaking): Functions that can fail now return LiteRtLmStatusCode with results in out_* parameters, and only litert_lm_* symbols are exported.
- ⚙️ Performance & Platform: GPU sampling now matches CPU for the same seed, ModelInfo is faster and uses less memory.