Skip to content

HandLandmarker killed by internal 'drishti' worker thread (SIGKILL) on Arch Linux #6375

Open

Description

@SidBroGG

Have I written custom code (as opposed to using a stock example script provided in MediaPipe)

Yes

OS Platform and Distribution

Arch Linux (Kernel: 7.2.7-arch1-1) - x86_64, Wayland

Mobile device if the issue happens on mobile device

No response

Browser and version if the issue happens on browser

No response

Programming Language and version

Python 3.14

MediaPipe version

1.0.1-1 python-mediapipe-bin aur

Bazel version

No response

Solution

HandLandmarker

Android Studio, NDK, SDK versions (if issue is related to building in Android environment)

No response

Xcode & Tulsi version (if issue is related to building for iOS)

No response

Describe the actual behavior

The Python process is abruptly terminated with PID killed python main.py (signal 9 / SIGKILL) right after initializing HandLandmarker options and creating the TFLite XNNPACK delegate. Tracking signal generation via bpftrace reveals that the kill signal is not emitted by the Linux kernel OOM killer or external system daemons, but originates internally from MediaPipe's own worker thread (drishti). The issue does not occur on Windows with identical pipeline logic and models, but fails consistently on Arch Linux regardless of whether packages are installed via yay (python-mediapipe-bin) or pip.

Describe the expected behaviour

HandLandmarker.create_from_options(options) should complete initialization and process video frames asynchronously via detect_async in LIVE_STREAM mode without triggering an internal panic/SIGKILL.

Standalone code/steps you may have used to try to get what you need

Install packages via yay: yay -S python-mediapipe-bin python-opencv python-numpy. Download model bundle: wget -O hand_landmarker.task https://storage.googleapis.com/mediapipe-models/hand_landmarker/hand_landmarker/float16/latest/hand_landmarker.task. Run script:

import time
import threading
from typing import Optional, List, Tuple
import cv2
import numpy as np
import mediapipe as mp
from mediapipe.tasks import python
from mediapipe.tasks.python import vision


class HandTracker:
    HAND_CONNECTIONS = [
        (0, 1), (1, 2), (2, 3), (3, 4),
        (0, 5), (5, 6), (6, 7), (7, 8),
        (5, 9), (9, 10), (10, 11), (11, 12),
        (9, 13), (13, 14), (14, 15), (15, 16),
        (13, 17), (17, 18), (18, 19), (19, 20),
        (0, 17)
    ]

    def __init__(self, model_path: str = "hand_landmarker.task", num_hands: int = 1):
        self._model_path = model_path
        self._num_hands = num_hands
        self._landmarker: Optional[vision.HandLandmarker] = None

        self._lock = threading.Lock()
        self._latest_result: Optional[vision.HandLandmarkerResult] = None
        self._last_timestamp_ms: int = 0

    def _result_callback(self, result: vision.HandLandmarkerResult, output_image: mp.Image, timestamp_ms: int):
        with self._lock:
            self._latest_result = result

    def start(self):
        options = vision.HandLandmarkerOptions(
            base_options=python.BaseOptions(model_asset_path=self._model_path),
            running_mode=vision.RunningMode.LIVE_STREAM,
            num_hands=self._num_hands,
            min_hand_detection_confidence=0.5,
            min_hand_presence_confidence=0.5,
            min_tracking_confidence=0.5,
            result_callback=self._result_callback
        )
        self._landmarker = vision.HandLandmarker.create_from_options(options)
        return self

    def close(self):
        if self._landmarker:
            self._landmarker.close()
            self._landmarker = None

    def __enter__(self):
        return self.start()

    def __exit__(self, exc_type, exc_val, exc_tb):
        self.close()

    def process_frame_async(self, bgr_frame: np.ndarray):
        if self._landmarker is None:
            raise RuntimeError("HandTracker is not running.")

        rgb_frame = cv2.cvtColor(bgr_frame, cv2.COLOR_BGR2RGB)
        mp_image = mp.Image(image_format=mp.ImageFormat.SRGB, data=rgb_frame)

        current_timestamp_ms = int(time.time() * 1000)
        if current_timestamp_ms <= self._last_timestamp_ms:
            current_timestamp_ms = self._last_timestamp_ms + 1
        self._last_timestamp_ms = current_timestamp_ms

        self._landmarker.detect_async(mp_image, current_timestamp_ms)

    def get_landmarks_pixel(self, frame_shape: Tuple[int, int, ...]) -> Optional[List[Tuple[int, int]]]:
        h, w = frame_shape[:2]
        with self._lock:
            if not self._latest_result or not self._latest_result.hand_landmarks:
                return None

            first_hand = self._latest_result.hand_landmarks[0]
            return [(int(pt.x * w), int(pt.y * h)) for pt in first_hand]


def draw_hand(frame: np.ndarray, landmarks: List[Tuple[int, int]]):
    for start_idx, end_idx in HandTracker.HAND_CONNECTIONS:
        p1 = landmarks[start_idx]
        p2 = landmarks[end_idx]
        cv2.line(frame, p1, p2, (0, 255, 0), 2)

    for pt in landmarks:
        cv2.circle(frame, pt, 4, (0, 0, 255), -1)


def main():
    cap = cv2.VideoCapture(0)
    if not cap.isOpened():
        print("Error: Could not open camera.")
        return

    with HandTracker(model_path="hand_landmarker.task", num_hands=1) as tracker:
        while True:
            ret, frame = cap.read()
            if not ret:
                break

            frame = cv2.flip(frame, 1)
            tracker.process_frame_async(frame)
            landmarks = tracker.get_landmarks_pixel(frame.shape)

            if landmarks:
                draw_hand(frame, landmarks)
                index_tip = landmarks[8]
                cv2.circle(frame, index_tip, 8, (255, 255, 0), -1)

            cv2.imshow("MediaPipe Tracker", frame)
            if cv2.waitKey(1) & 0xFF == ord('q'):
                break

    cap.release()
    cv2.destroyAllWindows()


if __name__ == "__main__":
    main()

Other info / Complete Logs

INFO: Created TensorFlow Lite XNNPACK delegate for CPU.
[1]    55326 killed      python main.py

Kernel log (dmesg) and systemd-oomd confirm no system OOM events occurred:

$ sudo dmesg -T | grep -Ei "killed process|oom"
[Tue Oct  6 14:57:13 2026] OOM killer disabled.
[Tue Oct  6 14:57:13 2026] OOM killer enabled.

$ free -h
               total        used        free      shared  buff/cache   available
Mem:            11Gi       3.3Gi       3.6Gi        71Mi       4.3Gi       8.1Gi
Swap:             0B          0B          0B

BpfTrace probe (tracepoint:signal:signal_generate /args->sig == 9/) output:

SENDER: drishti (PID 57337) -> VICTIM: PID 57376

strace inspection shows initialization proceeds normally through mmap of hand_landmarker.task, spawning the thread pool (ThreadPoolExecu), setting thread name to drishti, configuring timers (timer_create(..., SIGPROF)), after which all spawned threads simultaneously receive SIGKILL without any SIGSEGV or SIGABRT being logged first.

Activity

  1. added
    type:bugBug in the Source Code of MediaPipe Solution
    on Oct 6, 2026
  2. HarishPermal commented on Oct 7, 2026

    @HarishPermal

    hi @SidBroGG

    Thanks for reporting, i'm trying to reproduce this issue and will get back with update.

  3. HarishPermal commented on Oct 7, 2026

    @HarishPermal

    Hi @SidBroGG,

    I tested this in a clean Arch Linux docker container using Python 3.14.7, the official MediaPipe 1.1.0 wheel, and a 512 MiB memory limit.

    user@ubuntu:~/Documents/mediapipe_issues/6375$ docker run --rm -it \
      --memory=512m \
      --memory-swap=512m \
      --entrypoint /bin/bash \
      mediapipe-6375-arch
    [root@e4c254bba687 /]# python --version
    Python 3.14.7
    [root@e4c254bba687 /]# /opt/venv/bin/python -c \
      'import mediapipe as mp; print(mp.__version__, mp.__file__)'
    1.1.0 /opt/venv/lib/python3.14/site-packages/mediapipe/__init__.py
    [root@e4c254bba687 /]# /opt/venv/bin/python \
      /opt/repro/reproduce_issue.py \
      /opt/repro/hand_landmarker.task
    Python: 3.14.7 (main, Aug 10 2026, 07:46:56) [GCC 16.1.1 20260728]
    Platform: Linux-7.1.6-1rodete1-amd64-x86_64-with-glibc2.44
    MediaPipe: 1.1.0
    INFO: Created TensorFlow Lite XNNPACK delegate for CPU.
    WARNING: Logging before InitGoogle() is written to STDERR
    W0000 00:00:1791346282.185123      57 inference_feedback_manager.cc:121] Feedback manager requires a model with a single signature inference. Disabling support for feedback tensors.
    W0000 00:00:1791346282.209802      59 inference_feedback_manager.cc:121] Feedback manager requires a model with a single signature inference. Disabling support for feedback tensors.
    HandLandmarker initialized successfully

    Could you please retry using MediaPipe 1.1.0 in a clean virtual environment?

    python3.14 -m venv .venv
    .venv/bin/python -m pip install --upgrade pip
    .venv/bin/python -m pip install "mediapipe>=1.1.0"
    
    .venv/bin/python -c \
      'import mediapipe as mp; print(mp.__version__, mp.__file__)'
    
  4. SidBroGG commented on Oct 7, 2026

    @SidBroGG
    Author

    Hi, I tried running the code in a virtual environment, but the issue still persists

    ➜  mediapipe-issue python3.14 -m venv .venv
    .venv/bin/python -m pip install --upgrade pip
    .venv/bin/python -m pip install "mediapipe>=1.1.0"
    
    .venv/bin/python -c \
      'import mediapipe as mp; print(mp.__version__, mp.__file__)'
    Requirement already satisfied: pip in ./.venv/lib/python3.14/site-packages (26.2.1)
    Collecting mediapipe>=1.1.0
      Downloading mediapipe-1.1.0-py3-none-manylinux_2_28_x86_64.whl.metadata (10 kB)
    Collecting absl-py~=2.3 (from mediapipe>=1.1.0)
      Using cached absl_py-2.5.0-py3-none-any.whl.metadata (3.3 kB)
    Collecting certifi (from mediapipe>=1.1.0)
      Using cached certifi-2026.7.22-py3-none-any.whl.metadata (2.5 kB)
    Collecting numpy (from mediapipe>=1.1.0)
      Using cached numpy-2.5.3-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (6.6 kB)
    Collecting sounddevice~=0.5 (from mediapipe>=1.1.0)
      Using cached sounddevice-0.5.6-py3-none-any.whl.metadata (1.4 kB)
    Collecting flatbuffers~=25.9 (from mediapipe>=1.1.0)
      Using cached flatbuffers-25.12.19-py2.py3-none-any.whl.metadata (1.0 kB)
    Collecting opencv-contrib-python (from mediapipe>=1.1.0)
      Using cached opencv_contrib_python-5.0.0.93-cp37-abi3-manylinux_2_28_x86_64.whl.metadata (19 kB)
    Collecting matplotlib (from mediapipe>=1.1.0)
      Downloading matplotlib-3.11.2-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (80 kB)
    Collecting cffi (from sounddevice~=0.5->mediapipe>=1.1.0)
      Using cached cffi-2.1.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (2.5 kB)
    Collecting pycparser (from cffi->sounddevice~=0.5->mediapipe>=1.1.0)
      Using cached pycparser-3.0-py3-none-any.whl.metadata (8.2 kB)
    Collecting contourpy>=1.0.1 (from matplotlib->mediapipe>=1.1.0)
      Downloading contourpy-1.4.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (3.8 kB)
    Collecting cycler>=0.10 (from matplotlib->mediapipe>=1.1.0)
      Downloading cycler-0.12.1-py3-none-any.whl.metadata (3.8 kB)
    Collecting fonttools>=4.28.2 (from matplotlib->mediapipe>=1.1.0)
      Downloading fonttools-4.66.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (130 kB)
    Collecting kiwisolver>=1.3.1 (from matplotlib->mediapipe>=1.1.0)
      Downloading kiwisolver-1.5.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl.metadata (5.2 kB)
    Collecting packaging>=20.0 (from matplotlib->mediapipe>=1.1.0)
      Downloading packaging-26.3-py3-none-any.whl.metadata (3.5 kB)
    Collecting pillow>=9 (from matplotlib->mediapipe>=1.1.0)
      Downloading pillow-12.3.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl.metadata (9.1 kB)
    Collecting pyparsing>=3 (from matplotlib->mediapipe>=1.1.0)
      Downloading pyparsing-3.3.3-py3-none-any.whl.metadata (5.9 kB)
    Collecting python-dateutil>=2.7 (from matplotlib->mediapipe>=1.1.0)
      Downloading python_dateutil-2.9.0.post0-py2.py3-none-any.whl.metadata (8.4 kB)
    Collecting six>=1.5 (from python-dateutil>=2.7->matplotlib->mediapipe>=1.1.0)
      Using cached six-1.17.0-py2.py3-none-any.whl.metadata (1.7 kB)
    Downloading mediapipe-1.1.0-py3-none-manylinux_2_28_x86_64.whl (40.6 MB)
       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 40.6/40.6 MB 302.3 kB/s  0:02:01
    Using cached absl_py-2.5.0-py3-none-any.whl (137 kB)
    Using cached flatbuffers-25.12.19-py2.py3-none-any.whl (26 kB)
    Using cached sounddevice-0.5.6-py3-none-any.whl (32 kB)
    Using cached certifi-2026.7.22-py3-none-any.whl (136 kB)
    Using cached cffi-2.1.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (221 kB)
    Downloading matplotlib-3.11.2-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (11.0 MB)
       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 11.0/11.0 MB 620.5 kB/s  0:00:17
    Downloading contourpy-1.4.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (384 kB)
    Downloading cycler-0.12.1-py3-none-any.whl (8.3 kB)
    Downloading fonttools-4.66.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (5.3 MB)
       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 5.3/5.3 MB 203.7 kB/s  0:00:26
    Downloading kiwisolver-1.5.1-cp314-cp314-manylinux2014_x86_64.manylinux_2_17_x86_64.whl (1.5 MB)
       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 1.5/1.5 MB 388.7 kB/s  0:00:04
    Using cached numpy-2.5.3-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (16.7 MB)
    Downloading packaging-26.3-py3-none-any.whl (129 kB)
    Downloading pillow-12.3.0-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (6.9 MB)
       ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 6.9/6.9 MB 361.1 kB/s  0:00:18
    Downloading pyparsing-3.3.3-py3-none-any.whl (126 kB)
    Downloading python_dateutil-2.9.0.post0-py2.py3-none-any.whl (229 kB)
    Using cached six-1.17.0-py2.py3-none-any.whl (11 kB)
    Using cached opencv_contrib_python-5.0.0.93-cp37-abi3-manylinux_2_28_x86_64.whl (82.1 MB)
    Using cached pycparser-3.0-py3-none-any.whl (48 kB)
    Installing collected packages: flatbuffers, six, pyparsing, pycparser, pillow, packaging, numpy, kiwisolver, fonttools, cycler, certifi, absl-py, python-dateutil, opencv-contrib-python, contourpy, cffi, sounddevice, matplotlib, mediapipe
    Successfully installed absl-py-2.5.0 certifi-2026.7.22 cffi-2.1.1 contourpy-1.4.0 cycler-0.12.1 flatbuffers-25.12.19 fonttools-4.66.1 kiwisolver-1.5.1 matplotlib-3.11.2 mediapipe-1.1.0 numpy-2.5.3 opencv-contrib-python-5.0.0.93 packaging-26.3 pillow-12.3.0 pycparser-3.0 pyparsing-3.3.3 python-dateutil-2.9.0.post0 six-1.17.0 sounddevice-0.5.6
    1.1.0 /home/sidbro/local_repos/mediapipe-issue/.venv/lib/python3.14/site-packages/mediapipe/__init__.py
    ➜  mediapipe-issue .venv/bin/python -c \
      'import mediapipe as mp; print(mp.__version__, mp.__file__)'
    1.1.0 /home/sidbro/local_repos/mediapipe-issue/.venv/lib/python3.14/site-packages/mediapipe/__init__.py
    ➜  mediapipe-issue wget -O hand_landmarker.task https://storage.googleapis.com/mediapipe-models/hand_landmarker/hand_landmarker/float16/latest/hand_landmarker.task
    --2026-10-07 09:16:29--  https://storage.googleapis.com/mediapipe-models/hand_landmarker/hand_landmarker/float16/latest/hand_landmarker.task
    Loaded CA certificate '/etc/ssl/certs/ca-certificates.crt'
    Resolving storage.googleapis.com (storage.googleapis.com)... 142.250.109.207, 142.251.98.207, 142.250.120.207, ...
    Connecting to storage.googleapis.com (storage.googleapis.com)|142.250.109.207|:443... connected.
    HTTP request sent, awaiting response... 200 OK
    Length: 7819105 (7.5M) [application/octet-stream]
    Saving to: ‘hand_landmarker.task’
    
    hand_landmarker.task               100%[==============================================================>]   7.46M   180KB/s    in 25s     
    
    2026-10-07 09:16:56 (304 KB/s) - ‘hand_landmarker.task’ saved [7819105/7819105]
    
    ➜  mediapipe-issue python main.py          
    WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
    I0000 00:00:1791353835.652098   91682 port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
    I0000 00:00:1791353835.704994   91682 cpu_feature_guard.cc:227] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
    To enable the following instructions: SSE3 SSE4.1 SSE4.2 AVX AVX2 AVX512F AVX512_VNNI AVX512_BF16 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
    WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
    I0000 00:00:1791353837.085394   91682 port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.
    [ WARN:[email protected]] global cap_gstreamer.cpp:1172 isPipelinePlaying OpenCV | GStreamer warning: GStreamer: pipeline have not been created
    [1]    91682 killed     python main.py
  5. SidBroGG commented on Oct 7, 2026

    @SidBroGG
    Author

    I just realized I ran main.py using the system python instead of the virtual environment's python. But again, the issue persists. Here is the updated log:

    ➜  mediapipe-issue python3.14 -m venv .venv
    .venv/bin/python -m pip install --upgrade pip
    .venv/bin/python -m pip install "mediapipe>=1.1.0"
    
    .venv/bin/python -c \
      'import mediapipe as mp; print(mp.__version__, mp.__file__)'
    Requirement already satisfied: pip in ./.venv/lib/python3.14/site-packages (26.2.1)
    Requirement already satisfied: mediapipe>=1.1.0 in ./.venv/lib/python3.14/site-packages (1.1.0)
    Requirement already satisfied: absl-py~=2.3 in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (2.5.0)
    Requirement already satisfied: certifi in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (2026.7.22)
    Requirement already satisfied: numpy in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (2.5.3)
    Requirement already satisfied: sounddevice~=0.5 in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (0.5.6)
    Requirement already satisfied: flatbuffers~=25.9 in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (25.12.19)
    Requirement already satisfied: opencv-contrib-python in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (5.0.0.93)
    Requirement already satisfied: matplotlib in ./.venv/lib/python3.14/site-packages (from mediapipe>=1.1.0) (3.11.2)
    Requirement already satisfied: cffi in ./.venv/lib/python3.14/site-packages (from sounddevice~=0.5->mediapipe>=1.1.0) (2.1.1)
    Requirement already satisfied: pycparser in ./.venv/lib/python3.14/site-packages (from cffi->sounddevice~=0.5->mediapipe>=1.1.0) (3.0)
    Requirement already satisfied: contourpy>=1.0.1 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (1.4.0)
    Requirement already satisfied: cycler>=0.10 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (0.12.1)
    Requirement already satisfied: fonttools>=4.28.2 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (4.66.1)
    Requirement already satisfied: kiwisolver>=1.3.1 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (1.5.1)
    Requirement already satisfied: packaging>=20.0 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (26.3)
    Requirement already satisfied: pillow>=9 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (12.3.0)
    Requirement already satisfied: pyparsing>=3 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (3.3.3)
    Requirement already satisfied: python-dateutil>=2.7 in ./.venv/lib/python3.14/site-packages (from matplotlib->mediapipe>=1.1.0) (2.9.0.post0)
    Requirement already satisfied: six>=1.5 in ./.venv/lib/python3.14/site-packages (from python-dateutil>=2.7->matplotlib->mediapipe>=1.1.0) (1.17.0)
    1.1.0 /home/sidbro/local_repos/mediapipe-issue/.venv/lib/python3.14/site-packages/mediapipe/__init__.py
    ➜  mediapipe-issue .venv/bin/python main.py
    [1]    180676 killed     .venv/bin/python main.py
  6. HarishPermal commented on Oct 12, 2026

    @HarishPermal

    Thanks @SidBroGG let me test again, i'll update you

  7. KacperPiotrowski commented on Oct 18, 2026

    @KacperPiotrowski

    Hi MediaPipe team!

    We may have another reproduction of the SIGKILL / exit code 137 problem described here, but this time with FaceLandmarker rather than HandLandmarker.

    Environment

    • Arch Linux, KDE/Wayland, x86_64
    • Python 3.14 in an isolated virtual environment
    • Official PyPI mediapipe==1.1.1 wheel
    • Official face_landmarker.task model
    • 64 GB RAM, approximately 53 GiB available during diagnostics
    • No OOM or systemd-oomd events found in the inspected logs

    Minimal reproduction

    from pathlib import Path
    import mediapipe as mp
    
    print("A: MediaPipe imported", flush=True)
    
    options = mp.tasks.vision.FaceLandmarkerOptions(
        base_options=mp.tasks.BaseOptions(
            model_asset_path=str(Path("models/face_landmarker.task").resolve())
        ),
        running_mode=mp.tasks.vision.RunningMode.VIDEO,
        num_faces=1,
        output_face_blendshapes=False,
    )
    
    print("B: Initializing...", flush=True)
    
    with mp.tasks.vision.FaceLandmarker.create_from_options(options):
        print("C: Initialized successfully", flush=True)

    Actual result

    A: MediaPipe imported
    B: Initializing...
    WARNING: Logging before InitGoogle() is written to STDERR
    W0000 ... face_landmarker_graph.cc:180] Sets FaceBlendshapesGraph acceleration to xnnpack by default.
    Killed
    Exit status: 137
    

    The process is killed during FaceLandmarker.create_from_options(), before processing any images.

    We reproduced this both with and without an OpenCV GUI, and with output_face_blendshapes=False. The standalone reproducer does not access any camera or other hardware.

    This seems related to the original HandLandmarker report, but we have not traced the source of SIGKILL, so we cannot yet confirm the same root cause.

    Our downstream investigation:
    KacperPiotrowski/Kinect2Linux#25

    Could you advise which additional diagnostics would be most useful to identify the source of SIGKILL on MediaPipe 1.1.1?

    Thank you!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

platform:pythonMediaPipe Python issuestype:bugBug in the Source Code of MediaPipe Solution

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions