Owner

google-ai-edge

10 indexed repositories · View on GitHub

mediapipe
Cross-platform, customizable ML solutions for live and streaming media.

C++ · 37006 stars
gallery
A gallery that showcases on-device ML/GenAI use cases and allows people to try and use models locally.

Kotlin · 24738 stars
LiteRT-LM
LiteRT-LM is Google's production-ready, high-performance, open-source inference framework for deploying Large Language Models on edge dev...

C++ · 6479 stars
LiteRT
LiteRT, successor to TensorFlow Lite. is Google's On-device framework for high-performance ML & GenAI deployment on edge platforms, via e...

C++ · 3418 stars
mediapipe-samples

Jupyter Notebook · 2831 stars
model-explorer
A modern model graph visualizer and debugger

JavaScript · 1562 stars
litert-torch
Support PyTorch model conversion with LiteRT.

Jupyter Notebook · 1098 stars
litert-samples
LiteRT and LiteRT-LM sample apps, model recipes, agent skills and utilities.

Python · 436 stars
ai-edge-quantizer
AI Edge Quantizer: flexible post training quantization for LiteRT models.

Python · 201 stars
LiteRT-CLI
A convenient CLI to streamline LiteRT related development workflows, including converting, quantizing, compiling, managing, running, benc...

Python · 42 stars

Beginner friendly Everything

Open All

0.18 nightly: a bundle whose chat template expects string content silently drops the user's prompt Open

Difficulty 4/5 3-5 days Newbie friendliness 58/100

google-ai-edge/LiteRT-LM#3688 · 1 comment · Opened Sep 20, 2026
Apps that use MediaPipe via Swift Package Manager cannot be uploaded to App Store Connect / TestFlight Open

type:build/install type:support

google-ai-edge/mediapipe#6367 · 1 assignee · Opened Sep 19, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#682 · 1 assignee · Opened Sep 19, 2026
[Tensor SDK][AVF] Device assignment of the Tensor TPU mailbox context (Pixel 11): will app-launched pVMs get it, and is Tensor G5 in scope? Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

google-ai-edge/LiteRT#10081 · Opened Sep 19, 2026
CompiledModel.from_file on a truncated .tflite ends the process with SIGSEGV (ai-edge-litert 2.2.0, macOS arm64); from_buffer and 2.1.6 raise Open

Difficulty 4/5 3-5 days Newbie friendliness 56/100

google-ai-edge/LiteRT#10080 · Opened Sep 19, 2026
omni/asr: moonshine-tiny (i8) on macOS Metal transcribes the first window and then drifts into hallucinated text (non-deterministic); the f32 file aborts in DelegatePrepare Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

google-ai-edge/LiteRT-LM#3685 · 1 reaction · Opened Sep 18, 2026
omni/asr: Qwen3-ASR-0.6B `.litertlm` returns an empty transcript with `--backend gpu` on macOS (Metal) and on a Galaxy S26 (OpenCL); the same file transcribes on CPU Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

google-ai-edge/LiteRT-LM#3684 · 1 reaction · Opened Sep 18, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#681 · 1 assignee · Opened Sep 18, 2026
[Feature Request] Expose run_text_scoring through the Kotlin/JNI binding Open

type: feature request

google-ai-edge/LiteRT-LM#3639 · 2 comments · 1 assignee · Opened Sep 17, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#680 · 1 assignee · Opened Sep 17, 2026
FunctionGemma: the published chat template and fc_tool_format_utils.cc render the same declarations differently, and the template does not reproduce the golden in function_gemma_data_processor_test.cc Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3638 · 1 comment · 2 assignees · Opened Sep 17, 2026
Install through Homebrew Open

platform:macOS type:feature

google-ai-edge/gallery#1215 · 1 comment · 1 assignee · Opened Sep 17, 2026
litert-torch 0.9.x silently produces wrong output for non-contiguous constants (correct on 0.8.0) Open

status:awaiting user response type:bug

google-ai-edge/litert-torch#1238 · 2 comments · 1 assignee · Opened Sep 16, 2026
Since the “token_count” display includes “reasoning” in its value, it is not a reliable measure of the token count. Open

status:awaiting user response type :bug

google-ai-edge/LiteRT-LM#3614 · 2 comments · 1 assignee · Opened Sep 16, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#679 · 1 assignee · Opened Sep 16, 2026
Add `py.typed` and type hints to Python package Open

type:feature

google-ai-edge/LiteRT#9973 · 1 comment · 1 assignee · Opened Sep 15, 2026
`dynamic_wi4b32_afp32`: near-zero blocks get float16 scales of 0 or subnormal, and the XNNPACK delegate then refuses the model (MiniCPM5-2B) Open

google-ai-edge/ai-edge-quantizer#678 · 2 assignees · Opened Sep 15, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#677 · 1 assignee · Opened Sep 15, 2026
WebGPU accelerator: `--enable_profiling` fails with "UNIMPLEMENTED: Profile is not implemented" (macOS, LiteRT-LM v0.16.0 and v0.17.0) Open

status:awaiting LiteRTer status:awaiting PR review type:accelerators type:bug type:gpu delegate

google-ai-edge/LiteRT#9967 · 1 comment · 3 assignees · Opened Sep 15, 2026
crash report Open

area:crashes platform:Android type:bug

google-ai-edge/gallery#1205 · 1 comment · 1 assignee · Opened Sep 14, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#676 · 1 assignee · Opened Sep 14, 2026
GPU (OpenCL, Adreno) output diverges from CPU for the MiniCPM5-2B `dynamic_wi4c_hr_afp32` bundle, on litertlm-android 0.17.0 and the v0.16.0 CLI Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3577 · 1 comment · 3 assignees · Opened Sep 13, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#675 · 1 assignee · Opened Sep 13, 2026
MiniCPM5-2B: CLI --preset tools stop at template render (tojson(ensure_ascii=False)), and the canonical MiniCPM 5 template's XML function call is not executed Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3576 · 1 comment · 3 assignees · Opened Sep 13, 2026
[Android] v0.17.0 OpenCL accelerator crashes on Galaxy S10 5G; adding libandroid.so to DT_NEEDED fixes it Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3575 · 1 comment · 2 assignees · Opened Sep 13, 2026
Add community C#/.NET bindings to the README Open

type: feature request

google-ai-edge/LiteRT-LM#3572 · 1 comment · 1 assignee · Opened Sep 12, 2026
Application window too large vertically for MacBook Air 15 inch Open

area:ui-bugs platform:macOS type:bug

google-ai-edge/gallery#1204 · 3 comments · 1 assignee · Opened Sep 12, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#674 · 1 assignee · Opened Sep 12, 2026
Text scoring rewinds the KV-cache position twice after prefill Open

status:awaiting LiteRT-LM developer status:awaiting PR review type :bug

google-ai-edge/LiteRT-LM#3561 · 2 comments · 2 assignees · Opened Sep 11, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#673 · 1 assignee · Opened Sep 11, 2026
Clarify Google Runtime Environment entry in v0.16.0 iOS SBOM Open

status:awaiting LiteRT-LM developer type:support

google-ai-edge/LiteRT-LM#3556 · 1 comment · 2 assignees · Opened Sep 11, 2026
The memory data printed by Tflite's benchmark is different from the actual test results Open

status:awaiting user response status:stale type:bug

google-ai-edge/LiteRT#9899 · 2 comments · 1 assignee · Opened Sep 11, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#672 · 1 assignee · Opened Sep 10, 2026
could one use version 1 of media pipe in a serial mode in cloud without delay in real time ? Open

os:linux-non-arm platform:python stat:awaiting response task:face landmarker

google-ai-edge/mediapipe#6365 · 2 comments · 1 assignee · Opened Sep 9, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#670 · 1 assignee · Opened Sep 9, 2026
Tasks Core/Vision 1.0.0 Any.Builder concrete-return calls versus declared protobuf-javalite dependency Open

platform:android type:build/install

google-ai-edge/mediapipe#6364 · 5 comments · 1 assignee · Opened Sep 8, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#669 · 1 assignee · Opened Sep 8, 2026
Android 0.16.1: exact native build provenance and GpuArtisan main-model dispatch/completion contract Open

status:awaiting LiteRT-LM developer type:support

google-ai-edge/LiteRT-LM#3512 · 1 comment · 2 assignees · Opened Sep 7, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#668 · 1 assignee · Opened Sep 7, 2026
Gemma-4-E2B-it crashes with Status 13 on GPU despite being marketed for mid-range hardware Open

area:crashes type:bug

google-ai-edge/gallery#1188 · 1 assignee · Opened Sep 7, 2026
Constrained decoding with a SentencePiece bundle: the llguidance token table gets the raw pieces, so constrained output can never contain a space; a regex that requires one never returns Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3510 · 2 comments · 2 assignees · Opened Sep 6, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#667 · 1 assignee · Opened Sep 6, 2026
[NPU] Gemma 3 bundles (270M and 1B) lose every prefill chunk after the first on Qualcomm and Intel NPU; on Qualcomm it tracks the prefill-mask size (>~1 MiB, cache_length >= 1024 at prefill 128) and cache_length 896 works; Gemma 4 is fine on both Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3508 · 10 comments · 2 assignees · Opened Sep 6, 2026
Android GPU: rearranged weights are copied to the GPU instead of mmap'd (iOS keeps mmap) — +780 MB resident vs iOS at the same 8192 window Open

status:awaiting LiteRT-LM developer type :bug

google-ai-edge/LiteRT-LM#3507 · 1 comment · 2 assignees · Opened Sep 5, 2026
Artifact-specific license and provenance for Face Landmarker Web bundle Open

stat:awaiting googler type:support

google-ai-edge/mediapipe#6355 · 2 comments · 2 assignees · Opened Sep 5, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#666 · 1 assignee · Opened Sep 5, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#665 · 1 assignee · Opened Sep 4, 2026
Add Signed Process Domain (SignedPD) support and fallback in HtpBackend::Init. Open

status:awaiting PR review type:feature

google-ai-edge/LiteRT#9754 · 2 comments · 1 assignee · Opened Sep 3, 2026
Colab Tests (nightly) GitHub Action is failing Open

bug

google-ai-edge/ai-edge-quantizer#664 · 1 assignee · Opened Sep 3, 2026
Provide ability to export a model with an MTP drafter section Open

status:awaiting ai-edge-developer type:support

google-ai-edge/litert-torch#1214 · 1 comment · 2 assignees · Opened Sep 2, 2026

Showing the newest 100

This page lists what was indexed most recently. Advanced filter has the whole inventory, narrowed by language, difficulty and how long a task takes.

Open advanced filter

google-ai-edge

Showing the newest 100

Get new issues in your inbox