Compreface gpu github Essentially, it is a docker-based application that can be used as a standalone server or Describe the bug SubCenter-ArcFace-r100-gpu docker cannot find face To Reproduce Steps to reproduce the behavior: docker pull exadel/compreface-core:0. yml at master · exadel-inc/CompreFace To be able to use a GPU enabled build in Docker Desktop Windows, we have to edit the docker-compose. Reload to refresh your session. Did I miss something? I agree with you, but in order to use CompreFace as an addon in HassOS, we need a single container, it's a pre-requirement in order to have a supported addon in the HA environment. 01 as per the posted compatible GPU's (not that I am trying to place blame - just stating it is supposed to be compatible). So e. PoseEstimator" Saved searches Use saved searches to filter your results more quickly CompreFace: - Supports both CPU and GPU and is easy to scale up - Is open source and self-hosted, which gives you additional guarantees for data security - Can be deployed either in the cloud or on premises - Can be set up and used without machine learning expertise - Uses FaceNet and InsightFace libraries, which use state-of-the-art face Exadel CompreFace is a free and open-source face recognition GitHub project. Hi All, I'm planning to use my server (i5-4690K, RAID, Ubuntu 22. ***> wrote: Did you change the base image inside the Dockerfile. That being said I do have to use older You can see from the image attached the CPU usage spike until i stop the compreface container, usage goes up to around 90%. How do I make a "proper compreface-core" image? Multiple gpu support. deepstack takes almost forever dont think thats a hardwarepoblem. For each process, it will take memory, both RAM and GPU. facemask. " Learn more Exadel CompreFace is a free and open-source face recognition GitHub project. Technology-wise, CompreFace has several advantages over similar free face recognition solutions. I'm not sure if I need to configure CompreFace to be local somehow. 0 while retaining the face collection data I just replaced the dockher-compose. 5GB RAM free which is not being used. I've started it by passing a specific device id in the docker-compose. I am using double-take to get compreface process the images. 1-arcface-r100-gpu to compreface:1. com/exadel CompreFace is delivered as a docker-compose config and supports different models that work on CPU and GPU. Memory usage shouldn't grow over time and previously processed images shouldn't mess up with the GPU's memory. Are the GPU versions too fast? Leading free and open-source face recognition system - exadel-inc/CompreFace CompreFace List of custom-builds. 0 Features Massive UX/UI redesign Wizard for creating the first facial recognition service Pose plugin Service statistics Automatic login after registration Updated session logic, no need to re-login GPU Support and information. Just use CompreFace kubernetes and report ideas and bugs on GitHub; Share knowledge and experience via posting guides and articles, or just improve our documentation; Add other kubernetes configs and tutorials. Get CompreFace on Github >> Features. yml and ran it with docker compose up. Understood. Do not stop it during this time, as it may corrupt database data during data migration. jpg images from Frigate's API. 04; GPU: GTX 970; Driver: 510. GenderDetector: gender: agegender: Tensorflow Leading free and open-source face recognition system - exadel-inc/CompreFace When the frigate/events topic is updated the API begins to process the snapshot. 108. Custom builds require CPU with AVX2 support. which works normally flawless. compreface. Automate any workflow Sign up for free to join this conversation on GitHub. CompreFace is running on Desktop Docker. OutOfMemoryError: Java heap space I have compreface application with few hundred thousand subjects. These images are passed from the API to the configured detector(s) until a match is found that meets the SDK supports all functionality from CompreFace. Already have an account? Description java. This computer has a NVIDIA RTX GPU. Wait a few minutes (~10 minutes in my last test) Expected behavior. GPU versions it just floods it. To Reproduce Steps to reproduce the behavior: These messages keep repeating in the logs: {"severity": "DEB machine: x86_64 clock source: unix detected number of CPU cores: 24 current working directory: /app/ml detected binary path: /usr/local/bin/uwsgi !!! no internal routing support, rebuild with pcre support !!! 2023-03-23 13:51:03,846 INFO success: compreface-core entered RUNNING state, process has stayed up for > than 0 seconds (startsecs i am running deepstack with gtx1660 and nvidia-gpu-container. 80%: InsightFace: Leading free and open-source face recognition system - exadel-inc/CompreFace We want to improve our open-source face recognition solution, so your contributions are welcome and greatly appreciated. You can run docker-compose ps to see all CompreFace services. CompreFace can be applied to any field, such as security, advertising, marketing, attendance, VIP services, as well as hotel, conference, and airport check-in. LandmarksDetector,agegender. maybe i misunderstood, but doubl-take uses "detectores" like compreface for face-detection. 8 and added support for the new generation of Nvidia GPUs, including the Ada Lovelace and Hopper microarchitectures. Interestingly the GPU still has over 1. than i changed to mobilenet a You signed in with another tab or window. the problem seems the suite itself compreface had better results, but both together slow down the whole doubletake. it would run best on. (With a ~4GB GPU to reproduce the conditions) Add faces 5 by 5. CompreFace: Supports both CPU and GPU and is easy to scale up; Is open source and self-hosted, which gives you additional guarantees for data security If yes - it should work with GPU with an index 0 To change this - you can set GPU_IDX environment variable But, I am not sure that the problem with compreface-core container. Leading free and open-source face recognition system - exadel-inc/CompreFace By default, the CompreFace release contains configuration that could be run on the widest variety of hardware. 03 Hi So I downloaded from the release the last version 1. 🟡 SDK works with this CompreFace version. Leading free and open-source face recognition system - OverTM/exadel-inc. 0 version Please check this article to make sure your GPU is supported. I don't have a clear answer on how to install CompreFace with GPU for Windows. 0. CompreFace: Supports both CPU and GPU and is easy to scale up; Is open source and self-hosted, which gives you additional guarantees for data security In the photo below, only 18 faces were recognized out of 32，use SubCenter-ArcFace-r100-gpu。 6、For the configurations compreface_api_java_options=-Xmx8g and compreface_admin_java_options=-Xmx8g, if the memory is increased, for example, to 16GB or 24GB, would it support a greater number of faces? Contribute to eyalfink/CompreFace development by creating an account on GitHub. so the detector must have an AI-Hardware for face-rec, right? but compreface dont need it, as mentioned in the docs , i found it interesting that it is possible to use GPU (as AI-Hardware) and more precise-facerec models with compreface. env file with configuration options for CompreFace. In case if CompreFace version is older - new SDK features will fail. Sign in Product Actions. 2 repository. Contribute to ninjasstudio/double-take-hassio-addons-GPU development by creating an account on GitHub. I see that it's possible to use the SubCenter-ArcFace-r100 custom build, but I don't underst Describe the bug Core occupies memory until eventually taking all memory available. Face identification – identify a person Saved searches Use saved searches to filter your results more quickly Saved searches Use saved searches to filter your results more quickly Saved searches Use saved searches to filter your results more quickly On Tue, 8 Nov 2022, 2:51 am Pospielov Serhii, ***@***. Setup: OS:Ubuntu 22. Right, that much I understand. ini Saved searches Use saved searches to filter your results more quickly Describe the bug detection hangs, because of processes killed. Still, I've heard that the precision of CPU and GPU may be different (like numbers after 0. For production systems, we recommend looking through them and set up CompreFace accordingly. MaskDetector,facenet. 4%: arcface-r100-msfdrop75 / 99. I know that it's not as easy a process as on Linux. In the release archive and all custom builds, there is a . @yeahme49 Awesome! Thanks for the great work. I'm currently using Windows 10, i7. if you have 6Gb GPU, you should set it to 1, if you have a 12Gb GPU you should set it to 2, If you have a 24Gb GPU, you should set it to 5. I mean, the models are the same in GPU and CPU versions. here is a part of the docker-compose file for building with a custom model with GPU support: ```yaml compreface-core: image: ${registry}compreface-core:${CORE_VERSION} container_name: Attaching to compreface-postgres-db, compreface-core, compreface-api, compreface-admin, compreface-ui compreface-admin | Listening for transport dt_socket at address: 5005 compreface-core | [uWSGI] getting INI configuration from uwsgi. the log scrolls up with some backtraces i cannot read To Reproduce Steps to reproduce the behavior: starting docker-compose going to webgui testing facerec, but it doesnt run Current use of runtime: nvidia in the compreface-core: service of docker-compose. E. Celeron CPU support AVX and AVX2 only from Tiger Lake generation (2020 year) Leading free and open-source face recognition system - prixaro/CompreFace-2024 Launch ArcFace GPU with only 1 process. Leading free and open-source face recognition system - CompreFace/docker-compose. On the other hand, the GPU will use more considerably more power compared to the highly efficient coral so if power bill is a concern then that could be a Theoretically yes. 2. Our solution is based on state-of-the-art methods and libraries like FaceNet CompreFace provides REST API for face recognition, face verification, face detection, landmark detection, age, and gender recognition and is easily deployed with docker. GenderDetector,facenet. Is it correct that it's using this much memory, and if so what is the minimum requirement to use it? Leading free and open-source face recognition system - Releases · exadel-inc/CompreFace Default version of CompreFace requires CPU with AVX support. 1. 0 and took the "docker-compose-gpu. 0 from 1. Any help would be greatly appreciated. In . yml and . What I know, is that NVIDIA GeForce GTX 650 Ti is not supported from CompreFace 1. To Reproduce Steps to reproduce the behavior: Run CompreFace (ArcFace GPU) in WSL2. . Whenever an images needs to be processed compreface-core shows the captio CompreFace 1. For release and pre-build images, it should be set to exadel/ value; postgres_username - username for Postgres database i tried to change another compre-face docker with a different detection-model than the default. For example the SubCenter-ArcFace-r100-gpu build:. g. But if I attempt to restart or down and up the docker. env there is a parameter uwsgi_processes that says how many processes to use. In case if CompreFace version is newer - SDK won't support new features of CompreFace. Again, works fine in compreface regular. You switched accounts on another tab or window. Essentially, it is a docker-based application that can be used as a standalone server or deployed in the cloud. registry - this is the docker hub registry. Our solution is based on state-of-the-art methods and libraries like FaceNet By default, the CompreFace release contains configuration that could be run on the widest variety of hardware. 04) with Frigate. You still can retrain this model by yourself and then build CompreFace using your model. AgeDetector,agegender. It would also support other models like compreface face detection. 1 version (Ampere Nvidia GPU generation is supported) Migration from 1. Add this topic to your repo To associate your repository with the compreface topic, visit your repo's landing page and select "manage topics. yml and I can indeed choose which GPU to run, but after that, that's it, I can only roll with one. lang. Leading free and open-source face recognition system - exadel-inc/CompreFace Updated CUDA to version 11. Could you send logs from compreface-core logs? If it's not working then probably you need more RAM. if the machine has 2 or more GPUs, what should I do to make the compreface parallel the work between them? Skip to content. Similar for me, I went from compreface:0. I have installed and successfully launched the Mobilenet-GPU custom build of compreface but for some reason it is not activating uwsgi on the GPU, rather using the CPU (see the logs). For 5000 faces probably you need about 8-10 Gb of RAM. Actually, I'm running on an unRaid host I had to quickly use docker exec <container_name> mkdir -p /yolo4 while I started the container from the unRaid UI. Hi, for some reason my compreface docker doesn't work anymore. am running facenet with GPU support - insightface wouldnt run, not sure why but will try again later. CompreFace can be applied to any field, such as security, advertising, CompreFace is delivered as a docker-compose config and supports different models that work on CPU and GPU. ARG EXTRA_PLUGINS="facenet. Hi ! For some reason, whenever I start compreface, no matter the values of uwsgi_processes and uwsgi_threads, the processes are all started on the first available GPU. gpu file? I mean, when you run any command like RUN apt-key, it should depend not on your host OS but on OS inside docker. The downside of this build is that it's not optimized for the latest generations of Support CPU without AVX2. So Leading free and open-source face recognition system - exadel-inc/CompreFace On page load of /config or every 30 seconds, the detectors status is updated. There shouldn't be any spamming of the CompreFace API unless you keep refreshing the /config page on the DT UI. In one of the threads, you asked about adding processes in Python. env files in my custom-build "Subcenter-arcface-r100-gpu" directory by downloading these files from 1. Contribute to tyabru/CompreFace development by creating an account on GitHub. CompreFace Update compreface to 1. Template is the GPU arcface 1. I think the problem can be with compreface-api container. SDK supports all functionality from CompreFace. Original: compreface-core: image: ${registry}compreface-core:${CORE_VERSION} restart: always container_name: "compreface-core" runtime: nvidia environment: - ML_PORT=3000 - Plugin name Slug Backend Framework GPU support; agegender. There should be 5 Hi all, I'm not very happy with the results of the default installation of Compreface. Got it working using your docker image but with one small niggle: I had to manually create the directory /yolo4 before the container would start. 1-arcface-r100-gpu docker run -dp 3000:3000 - After you run CompreFace, wait at least 30 seconds until it starts. io Add-ons Compreface GPU. New User - Coral TPU vs TensorRT GPU. Features Updated CUDA to 11. So I would recommend getting GPU with at least 6Gb of memory. I am running this with Frigate and Double Take. I've got lots of false positives, with a very high score. Consume the API for 4-5h (sometimes less). All reactions Describe the bug CompreFace is slowly using more and more RAM. So the task is still to make a proper compreface-core image so that you can build a single image later. Those large collection applic You send logs only from compreface-api node. train false positives or tell CompreFace that it got the recognition wrong. How can i check if it is using the GPU as surely with the spike in CPU usage which is also causing CPU temps to reach over 70C shows that its not running on the GPU? i have the same problem, config two processs and one thread, the GPU memory only increases sometimes. CompreFace is delivered as a docker-compose config and supports different models that work on CPU and GPU. yml file of the build. Now there are approximately 8K persons added and the face recognition simply returns: [message] => Something went wrong, I bought this GPU based on the fact that it works with compreface 1. If you use the CPU versi Leading free and open-source face recognition system - Releases · exadel-inc/CompreFace CompreFace is delivered as a docker-compose config and supports different models that work on CPU and GPU. 0-arcface-r100-gpu and many pictures where faces were recognised and identified before now shows as "no face" Sure, many of them are a bit darker, or the angle of the face is not 100% straight forward. Expected behavior Continuo This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository. Exadel CompreFace is a free and open-source face recognition GitHub project. i tried the hardest-one without GPU, and the system stood nearly still. Describe the solution you'd like Double Take Hass. Navigation Menu Toggle navigation. jpg and latest. master You signed in with another tab or window. yml" and changed the name to docker-compose. AgeDetector: age: agegender: Tensorflow: agegender. 0000000001, for example). Currently have it running on a old iMac which seems to work fine (intel etc) but anything with a GPU in it would be even better. Probably we can do something to simplify this process. This is technically impossible, we take the face recognition model and use it for all recognitions. And when I checked, it looks like they really work the same. Until here all working well When I check the nvidia-smi with w To build CompreFace with the possibility to run without AVX, you need to compile Tensorflow without AVX support and then put it in the dockerfile of the 'compreface-core' image. OS: Unraid Custom Build: Mobilenet-GPU GPU: NVIDIA Quadro P5000 Describe the bug When I had a few persons ( subjects) stored in Compreface, the facial recognition worked just fine. yml for GPU enabled builds means that they will not work on Docker Desktop Windows (At least Is your feature request related to a problem? Single container build uses compreface-core image as a base and then adds everything else inside. Here is the CompreFace repository: https://github. 6. Added support for the Pose Plugin in the user interface. The number of The system can accurately identify people even when it has only “seen” their photo once. Custom-build Base library CPU GPU Face detection model / accuracy on WIDER Face (Hard) Face recognition model / accuracy on LFW SubCenter-ArcFace-r100-gpu: InsightFace: x86 (AVX2 instructions) GPU (CUDA required) retinaface_r50_v1 / 91. 0 version and ran the commands "docker compose down" and then "docker compose up -d" Sign up for free to join Python SDK for CompreFace - free and open-source face recognition system from Exadel - Workflow runs · exadel-inc/compreface-python-sdk Exadel CompreFace is a free and open-source face recognition GitHub project. You signed out in another tab or window. How can i enable this instead of using the CPU? I've tried following all online docs relating to the WSL2 but when processing images, it is still using the CPU not the GPU. could you run docker ps and check if it's running and run docker logs compreface-api The system can accurately identify people even when it has only “seen” their photo once. Leading free and open-source face recognition system - Workflow runs · exadel-inc/CompreFace first option worked - took a bit of messing around in docker renaming the container and getting it into the correct host network for the URL from within compreface-ui to resolve to compreface-core properly, but its running successfully and reporting available services so far. You don’t need prior machine learning skills to set up and use CompreFace. noern llfsbb klidhy lljwia qazurz telj avickvx ifsbto ukinqfq ywt