Repository Issue Activity (beta)

#1188 [Feature Request] Public API for LoRA adapter weights in Sessions

#2507 Feature Request: Enable MediaTek NPU acceleration (Dimensity 9300+ / MT6989) via LiteRT

#1662 Kotlin support for constrained decoding

#2506 Add an option to toggle thinking mode in the CLI app

#2505 Add an option to use custom number of CPU threads in the CLI app

#2504 iOS Simulator (arm64): CPU inference fails INTERNAL at llm_litert_compiled_model_executor.cc:755; GPU 'Failed to create engine'

#2503 MTP + GPU (Metal): mldrift cache files corrupted on 2nd+ initialization (iOS)

#2180 v0.11.0 Crashing on gpu backend with Intel A770 xpu

#2073 libLiteRtTopKWebGpuSampler exports only 3/7 C ABI functions on macOS/Windows (Linux/Android ship 7) — sampler_factory falls back to CPU sampling

#2502 Metal executor doesn't wire LiteRtTopKMetalSampler_SetInputTensorsAndInferenceFunc — on-GPU sampling unreachable on macOS at v0.12.0

#2497 Clarify Gemma 4 QAT support path for LiteRT-LM packages

#1862 Whenever I double-click an installer on Win11, this error pops up immediately

#2500 ADMINISTRATIVE NOTICE: SYSTEMIC INTELLECTUAL PROPERTY OVERLAP, COPYRIGHT INFRINGEMENT, AND PROVENANCE CONTAMINATION REGARDING GEMMA 4 QAT (wNa8o8 FORMAT)

#2499 OpenCL shader compilation error: convert_float(half4) overload resolution fails on Windows / NVIDIA RTX 2080 Ti (v0.9.0-alpha03)

#2461 Gemma 4 12B LiteRT-LM model fails to load on Android with LiteRT-LM 0.13.0

#2498 Clarify Gemma 4 12B audio MTP support path for LiteRT-LM packages

#2451 Gemma4 model support for VPUX37XX (ArrowLake NPU 3720) in litert-lm backend=npu

#2472 [Performance Regression] GPU decode speed drops to ~1/3 starting v0.13.0 (for GPU-embedding models)

#2480 Extreme RAM Staging Overhead (GTT Mirroring) on Linux with Discrete GPUs

#2479 buffer moved twice

#2442 Missing: `libclang_rt.profile-x86_64.a` : tip on Debian

#2443 Missing: `libclang_rt.profile-x86_64.a` : tip on Debian

#2424 Feature request: Support thinking budget for reasoning models

#2436 https://google-ai-edge.github.io/LiteRT-LM/ is 404

#2471 openai_server: tool_choice ignored — required, none, and named-tool all behave as auto