google-ai-edge/mediapipe issues and pull requests

#5533 - There is not enough memory left to execute the command

Issue - State: open - Opened by mmaciola 5 months ago - 8 comments
Labels: platform:android, stat:awaiting response, type:bug, task:face detection

#5530 - Differences and usage of outputSegmentationMasks optiom when using

Issue - State: open - Opened by theguaz 5 months ago - 2 comments
Labels: platform:javascript, type:bug, task:pose landmarker, os:macOS

#5521 - Added iOS audio classifier implementation

Pull Request - State: closed - Opened by priankakariatyml 5 months ago

#5520 - How to smooth the jitter from pose landmarker model?

Issue - State: open - Opened by ly15927086342 5 months ago - 1 comment
Labels: type:support, platform:javascript, task:pose landmarker, os:macOS

#5519 - Is it possible to manually map face landmarks for images where mediapipe doesn't detect them at all?

Issue - State: open - Opened by aesanchezgh 5 months ago - 2 comments
Labels: type:support, task:face landmarker

#5518 - Mediapipe LLM build fails for Android X86_64

Issue - State: closed - Opened by vraghavulu 5 months ago - 6 comments
Labels: type:build/install, platform:android, platform:c++, os:linux-non-arm, task:LLM inference

#5511 - 🐛 [mediapipe_task_genai] On Flutter/Android llm engine crashes : failed to validate graph

Issue - State: closed - Opened by tempo-riz 5 months ago - 5 comments
Labels: platform:android, type:bug, task:LLM inference

#5510 - mediapipe::CvPixelBufferPoolWrapper::GetBuffer() fatal: Check failed: !err Error creating pixel buffer: -6662

Issue - State: open - Opened by bhamiltoncx 5 months ago - 3 comments
Labels: type:bug

#5509 - Why does the mediapipe not use tflite runtime for genai?

Issue - State: closed - Opened by Hozzu 5 months ago - 5 comments
Labels: type:feature, platform:c++, os:linux-non-arm, task:LLM inference

#5503 - selfie_multiclass task with option delegate=BaseOptions.Delegate.GPU not working

Issue - State: open - Opened by SoaringTiger 5 months ago - 9 comments
Labels: gpu, platform:python, type:bug, stat:awaiting googler, platform:docker, task:image segmentation, os:linux-non-arm

#5486 - Make android solutions(face detection) working with 'minifyEnabled true' build option.

Issue - State: open - Opened by farmaker47 6 months ago - 5 comments
Labels: platform:android, type:bug, stat:awaiting googler, task:face detection

#5483 - unable to find stub files

Issue - State: closed - Opened by hermanumrao 6 months ago - 6 comments
Labels: type:feature, platform:python, legacy:pose

#5481 - module 'mediapipe.python._framework_bindings.model_ckpt_util' has no attribute 'ConvertHfTokenizer'

Issue - State: closed - Opened by mohamedlotfy50 6 months ago - 9 comments
Labels: os:windows, platform:python, type:bug, task:LLM inference

#5476 - NormalizedLandmark returns x,y,z coordinate values outside [0.0, 1.0]

Issue - State: open - Opened by ErikValle2 6 months ago - 2 comments
Labels: platform:python, type:bug, task:pose landmarker, os:linux-non-arm

#5468 - Pre-download mini-benchmark for in-browser (LLM) inference performance

Issue - State: closed - Opened by maudnals 6 months ago - 1 comment
Labels: type:feature, platform:javascript, task:LLM inference

#5462 - Mediapipe crashes with holistic model but not facemesh

Issue - State: closed - Opened by PaStMir 6 months ago - 5 comments
Labels: os:windows, platform:python, type:bug, task:holistic landmarker

#5459 - Crash when running Android text classification sample with MobileBERT.

Issue - State: closed - Opened by v-hogood 6 months ago - 11 comments
Labels: platform:android, type:bug, task:text classification

#5455 - After installing MediaPipeTasksVision with CocoaPods, I can no longer preview Views in SwiftUI

Issue - State: open - Opened by t-kashima 6 months ago - 4 comments
Labels: platform:ios, task::all, type:bug, stat:awaiting googler, os:macOS

#5452 - How to install and use GPU-supported MediaPipe on a Linux platform?

Issue - State: closed - Opened by caochengchen 6 months ago - 11 comments
Labels: type:support, gpu, task::all, platform:python, os:linux-non-arm

#5443 - Cannot run selfie segmentation in web worker - React.js

Issue - State: closed - Opened by DebasisDa 6 months ago - 7 comments
Labels: os:windows, platform:javascript, type:bug, legacy:selfie segmentation

#5441 - Landmark returning lagging by one frame in MediaPipe

Issue - State: closed - Opened by Tendaliu 6 months ago - 6 comments
Labels: os:windows, legacy:holistic, platform:c++, type:bug

#5415 - Updated the issue templates

Pull Request - State: closed - Opened by kuaashish 7 months ago - 2 comments

#5403 - Build error when adding mediapipe via cocoapods to the framework project (iOS)

Issue - State: open - Opened by kers97 7 months ago - 9 comments
Labels: platform:ios, type:build/install, task::all, stat:awaiting googler

#5399 - PyPI package updated without release notes, changelog, or Github release available

Issue - State: open - Opened by cchadowitz 7 months ago - 5 comments
Labels: type:others, type:docs-feature

#5394 - Issue with Pose Landmark Detection Segmentation Mask on iOS/iPadOS 17.4.1

Issue - State: open - Opened by cavisions 7 months ago - 2 comments
Labels: platform:ios, platform:c++, type:bug, stat:awaiting googler, task:pose landmarker

#5390 - Bump the npm_and_yarn group across 1 directory with 2 updates

Pull Request - State: closed - Opened by dependabot[bot] 7 months ago - 1 comment
Labels: dependencies

#5384 - Getting Buffer Validation newBufferWith must not exceed 256 MB error when I am trying to load model of 2.5 gb size

Issue - State: closed - Opened by Pratik-7i 7 months ago - 9 comments
Labels: platform:ios, stat:awaiting googler, task:LLM inference

#5372 - mediapipe 0.10.11 is wrongly depending on protobuf 3.x, it should be version > 4.x, even 5.x works

Issue - State: closed - Opened by alanwilter 7 months ago - 6 comments
Labels: platform:python, type:bug, task:face landmarker, os:macOS

#5371 - WARNING: All log messages before absl::InitializeLog() is called are written to STDERR

Issue - State: open - Opened by matanox 7 months ago - 9 comments
Labels: type:feature, platform:python, stat:awaiting googler, os:linux-non-arm

#5357 - Support for Converting Other LLMs like GPT-2 & QWEN in MediaPipe

Issue - State: closed - Opened by mayurmarvel 7 months ago - 9 comments
Labels: type:feature, platform:python, task:LLM inference

#5338 - No clear apks for quick testing

Issue - State: closed - Opened by StarShine1A 7 months ago - 5 comments
Labels: type:docs-bug, task::all, task:image generator, task:LLM inference

#5329 - When will the new, improved blendshape prediction model v3 be implemented into Mediapipe?

Issue - State: open - Opened by ferdiwerthi 8 months ago - 4 comments
Labels: type:feature, stat:awaiting googler, task:face landmarker

#5328 - Integrate MediaPipeTasksVision into dynamic framework throws Undefined symbols error

Issue - State: closed - Opened by zoltancsipler 8 months ago - 11 comments
Labels: platform:ios, type:build/install, task:face detection

#5327 - Allow cancelling ongoing async generation in LlmInference

Issue - State: closed - Opened by bringert 8 months ago - 3 comments
Labels: type:feature, platform:android, stat:awaiting googler, task:LLM inference

#5326 - <uses-native-library> missing in Android LLM Inference Guide

Issue - State: closed - Opened by bringert 8 months ago - 7 comments
Labels: type:docs-bug, platform:android, task:LLM inference

#5306 - You may need an additional loader to handle the result of these loaders.

Issue - State: closed - Opened by skandula106 8 months ago - 6 comments
Labels: task::all, os:windows, platform:javascript

#5305 - Add GPU support to LLM C API

Issue - State: closed - Opened by huazhusun 8 months ago - 4 comments
Labels: type:feature, platform:android, task:LLM inference

#5302 - iOS: Can't merge user_target_xcconfig for pod targets: ["MediaPipeTasksCommon", "MediaPipeTasksGenAIC"].

Issue - State: closed - Opened by DarrylBayliss 8 months ago - 8 comments
Labels: type:feature, platform:ios, task::all, task:LLM inference

#5295 - Ability to fine-tune pose landmark detection model with model maker

Issue - State: open - Opened by RoadieHouse 8 months ago - 7 comments
Labels: type:feature, platform:python, stat:awaiting googler, type:modelmaker, task:pose landmarker

#5288 - Fixed broken links. (issue: #5286)

Pull Request - State: open - Opened by AnirudhKaranth 8 months ago - 1 comment

#5284 - CanvasRenderingContext2D is not defined in DrawingUtils constructor

Issue - State: open - Opened by ddgg-el 8 months ago - 2 comments
Labels: type:support, platform:javascript, stat:awaiting googler, task:pose landmarker, os:macOS

#5277 - How to set a system prompt for RAG implementation for Inference for Gemma 2b on IOS ?

Issue - State: open - Opened by omkar806 8 months ago - 5 comments
Labels: type:feature, platform:ios, task:LLM inference

#5263 - MediaPipe GPU support in PyPi package

Issue - State: closed - Opened by JC3 8 months ago - 6 comments
Labels: type:feature, platform:python, task:pose landmarker

#5257 - Problem when using mediapipes from inside a worker loaded as a module

Issue - State: open - Opened by martenrichter 8 months ago - 14 comments
Labels: type:support, platform:javascript, stat:awaiting googler, task:image classification

#5255 - Merge LoRA weights to LLM at initialization time on-device (Gemma)

Issue - State: closed - Opened by sitatec 8 months ago - 3 comments
Labels: type:feature, platform:ios, platform:android, platform:javascript, task:LLM inference

#5215 - How can I Integrate NPU into

Issue - State: closed - Opened by aphlta 9 months ago - 7 comments
Labels: type:others

#5214 - Can't pip install mediapipe-model-maker

Issue - State: open - Opened by tomershukhman 9 months ago - 4 comments
Labels: platform:python, type:bug, stat:awaiting googler, type:modelmaker, os:macOS

#5212 - [HOLISTIC SOLUTION] Info about the visibility/confidence of keypoints from the hands is not available.

Issue - State: open - Opened by mvazquezgts 9 months ago - 8 comments
Labels: type:support, platform:python, stat:awaiting googler, os:linux-non-arm

#5201 - Need official proguard-rules.pro file for release build in Android

Issue - State: closed - Opened by albus-d 9 months ago - 9 comments
Labels: type:build/install, platform:android, task:audio classification

#5190 - The multi-class selfie segmenter does not run in GPU delegate mode on the Android Chrome browser.

Issue - State: closed - Opened by salopge 9 months ago - 24 comments
Labels: platform:android, platform:javascript, type:bug, task:image segmentation

#5181 - MediaPipe Holistic Landmarker in Python null_ptr exception

Issue - State: open - Opened by victoriayuechen 9 months ago - 24 comments
Labels: platform:python, type:bug, stat:awaiting googler, task:holistic landmarker

#5166 - 0.10.10 Holistic GPU delegation does not work

Issue - State: open - Opened by clemenshimmer 9 months ago - 14 comments
Labels: platform:javascript, type:bug, stat:awaiting googler, os:linux-non-arm, task:holistic landmarker

#5161 - Multi-person tracking of holistic model

Issue - State: closed - Opened by ChristianNSchmitz 9 months ago - 4 comments
Labels: type:feature, platform:python, task:pose landmarker

#5105 - errors encountered resolving select() keys for //mediapipe/framework:port

Issue - State: closed - Opened by KakaWan 10 months ago - 5 comments
Labels: type:build/install, os:windows

#5079 - Pose landmark detection Android code example crashes when startup on Samsung Galaxy S23 Ultra Android 14 OneUI6.0

Issue - State: open - Opened by at466MHz 11 months ago - 2 comments
Labels: type:support, platform:android, stat:awaiting googler, task:pose landmarker

#5065 - BlazePose Ghum Code and Model

Issue - State: open - Opened by kavikode 11 months ago - 3 comments
Labels: type:feature, platform:python, stat:awaiting googler, task:face landmarker

#5058 - Error on ios 14.6 safari：Check failed: prod_token

Issue - State: closed - Opened by huanhunx 11 months ago - 5 comments
Labels: platform:javascript, type:bug, task:face landmarker

#5027 - NPM package vision_bundle.mjs and .cjs file have incorrect sourceMappingURL

Issue - State: closed - Opened by samankittani 12 months ago - 5 comments
Labels: task::all, platform:javascript, type:bug, os:linux-non-arm

#5025 - Validation RECORD failed

Issue - State: open - Opened by alanwilter 12 months ago - 4 comments
Labels: task::all, platform:python, type:others, stat:awaiting googler

#5017 - The input packet MPImage returned by `onResult` after `detectAsync` call is corrupted

Issue - State: open - Opened by definability 12 months ago - 5 comments
Labels: platform:android, type:bug, stat:awaiting googler, task:hand landmarker

#5016 - Is body landmarks based segmentation mask no longer supported?

Issue - State: closed - Opened by Setsukyo 12 months ago - 10 comments
Labels: type:support, platform:python, task:pose landmarker

#4991 - A way to disable logging on mediapipe/tasks-vision

Issue - State: open - Opened by tomaszs about 1 year ago - 5 comments
Labels: type:feature, task::all, platform:javascript, stat:awaiting googler

#4990 - When BlazeFace (full-range) could be released?

Issue - State: open - Opened by StevenSK-king about 1 year ago - 9 comments
Labels: type:feature, platform:python, stat:awaiting googler, task:face detection

#4970 - Mediapipe python cuda gpu docker

Issue - State: closed - Opened by pydehon about 1 year ago - 10 comments
Labels: gpu, platform:python, type:others, platform:docker

#4958 - Logging issue when import mediapipe

Issue - State: open - Opened by 1019695938 about 1 year ago - 7 comments
Labels: stat:awaiting response, os:windows, platform:python, type:bug, os:linux-non-arm

#4916 - Error initializing PortAudio when started from Windows Service

Issue - State: open - Opened by jdlvwreygers about 1 year ago - 8 comments
Labels: os:windows, platform:python, type:bug, stat:awaiting googler, task:audio classification

#4895 - Documentation for building pose-tracking iOS app using mediapipe library with swift

Issue - State: open - Opened by diyoraharshitN about 1 year ago - 24 comments
Labels: platform:ios, stat:awaiting response, type:docs-feature, stale, task:pose landmarker

#4883 - Cross Compiling mediapipe on Ubuntu

Issue - State: closed - Opened by phil-fill about 1 year ago - 3 comments
Labels: type:build/install

#4778 - opencv2/core/version.hpp: No such file or directory

Issue - State: closed - Opened by VieiraJardel about 1 year ago - 6 comments
Labels: type:build/install

#4769 - Packet timestamp mismatch on a calculator receiving from stream "free_memory"

Issue - State: open - Opened by spennihana about 1 year ago - 6 comments
Labels: platform:javascript, type:bug, stat:awaiting googler, task:image segmentation, os:macOS

#4713 - How to turn off hand tracking in holistic?

Issue - State: open - Opened by justinduynguyen over 1 year ago - 4 comments
Labels: type:support, platform:javascript, legacy:holistic, stat:awaiting googler

#4540 - Error during Statically linking com.google.devtools.build.android.aapt2.CompiledResources@27912e3

Issue - State: closed - Opened by helloWorldNG over 1 year ago - 8 comments
Labels: type:build/install, platform:android, task::all

#4539 - Crash on run when loading from cache

Issue - State: closed - Opened by Perezmarc over 1 year ago - 5 comments
Labels: platform:javascript, type:bug, task:face detection

#4512 - Pip install mediapipe==0.10.0 not working !!!

Issue - State: closed - Opened by MH-Python over 1 year ago - 10 comments
Labels: type:build/install, task::all, platform:python

#4507 - Pose Landmarker Jittering

Issue - State: open - Opened by scottxp over 1 year ago - 37 comments
Labels: type:support, platform:javascript, stat:awaiting googler, task:pose landmarker

#4490 - The mapping between mediapipe and dlib's 68 key points

Issue - State: closed - Opened by shuishiwojiade over 1 year ago - 5 comments
Labels: type:others

#4423 - bject_detector.Dataset.from_pascal_voc_folder ignores negative/background images

Issue - State: open - Opened by willemnator over 1 year ago - 12 comments
Labels: platform:javascript, stat:awaiting googler, type:modelmaker, task:object detection

#4403 - tongueOut is missing in blendshapes output

Issue - State: open - Opened by jiangwei221 over 1 year ago - 7 comments
Labels: type:support, platform:python, type:bug, stat:awaiting googler, task:face landmarker

#4227 - Why is there no solution to run gpu mediapipe on windows 10?

Issue - State: closed - Opened by Mr-Khans over 1 year ago - 17 comments
Labels: type:build/install, gpu, os:windows

#4105 - MediaPipe Pose - mapping to SMPL model

Issue - State: closed - Opened by BrasD99 almost 2 years ago - 8 comments
Labels: type:support, platform:python, legacy:pose

#3837 - Update Dockerfile to install libopencv-contrib-dev

Pull Request - State: closed - Opened by plashchynski about 2 years ago

#3818 - mediapipe/holistic incorrect z-axis values for body pose landmarks

Issue - State: closed - Opened by vladmandic about 2 years ago - 12 comments
Labels: type:support, platform:javascript, legacy:holistic

#3505 - visibility and presence always be 0.

Issue - State: closed - Opened by endink over 2 years ago - 18 comments
Labels: type:support, task::all, legacy:holistic

#3499 - Pose 3D landmarks to Three.js

Issue - State: closed - Opened by ireneb612 over 2 years ago - 5 comments
Labels: type:support, legacy:hands, platform:javascript

#3400 - Facemesh.process() returns None after second image is sent to process.

Issue - State: closed - Opened by Kutayykilic over 2 years ago - 18 comments
Labels: type:support, legacy:face mesh, platform:python

#3189 - "Fail to read any frames from the video file"

Issue - State: closed - Opened by yanxky over 2 years ago - 9 comments
Labels: type:support, platform:desktop, legacy:iris

#3134 - GPU or CPU?

Issue - State: closed - Opened by lghasemzadeh over 2 years ago - 11 comments
Labels: type:support, gpu, task::all

#3042 - Mediapipe pose not working with react

Issue - State: closed - Opened by faraznisar almost 3 years ago - 8 comments
Labels: type:support, platform:javascript, legacy:pose

#3017 - INFO: Created TensorFlow Lite XNNPACK delegate for CPU

Issue - State: closed - Opened by akan7sha almost 3 years ago - 18 comments
Labels: type:support, legacy:hands, platform:python, stale

#2826 - Any solutions for Mixamo models translations

Issue - State: closed - Opened by shane99a about 3 years ago - 5 comments
Labels: legacy:pose, type:others

#2819 - How to clean up memory after model is not used anymore?

Issue - State: closed - Opened by ostap0207 about 3 years ago - 6 comments
Labels: type:support, platform:javascript, legacy:selfie segmentation

#2430 - error:could not con ＇result' from 'std: :unique_ptr <nediapipe: : {anonymous}::GlProcessor,std::default_delete<mediapi ｛anonymous} : : GLProcessor> >

Issue - State: closed - Opened by only-tao over 3 years ago - 10 comments
Labels: type:build/install, platform:python

GitHub / google-ai-edge/mediapipe issues and pull requests