Comfyui face detection not working. Open Cassady8912 opened this issue Oct 3 .
Comfyui face detection not working com). . gz (439 kB) Installing build dependencies: started Installing build dependencies: finished with status 'done' Getting requirements to build wheel: Welcome to the unofficial ComfyUI subreddit. exe if not working, try below. com/ltdrdata/ComfyUI-extension Stable Diffusion XL has trouble producing accurately proportioned faces when they are too small. [Feature]: Adding Face Recognition to the Face Detection Module to facilitate Face Swapping in videos with fast-moving crowds. I generally choose the second one. Currently, face detection using face indexes works exceptionally well, adhering to rules even in complex scenes or when faces vary in scale. As you can see in the preview image of the relevant portion of the workflow, just the body information seems to have been generated. This has spread as a tip for sharing model directories. This extension is particularly useful for AI artists looking to refine their portrait work or any project involving detailed face manipulation. but don't spam all your work. tar. Welcome to the unofficial ComfyUI subreddit. Face Detection Not Working: Make sure that the face detection feature is enabled in the Webcam app. I use paperspace and before starting comfyui I run !pip install insightface (and faceid works for me without a problem; check that in comfyui/models/ it is inside insightface and a folder called buffalo) Alternative try pip install onnxruntime Face detection models. I can't get this 896 x 1152 face-only Open pose to work with OpenPoseXL2. r/StableDiffusionInfo. You might need to hold your iPhone higher if you're wearing a face mask. 5 models sdxlfacedetail workflow. FaceDetailer does not support multi-batch, so you need to convert it to a list for processing. face module, OpenCV now recommends using the face_recognition library, which is a third-party library that provides a simple interface for face recognition tasks. Here's the links if you'd rather download them yourself. Others in the group are experiencing the same pr 0. Test Results: No, I say this because I have had compatibility problems with reactor and faceid. About Impact-Pack. bfloat16 Using pytorch cross attention Feature description Is it posssible to change detection model - for example to yolo8, or is there any folder to put it in pt. You'll find them here mav-rik/facerestore_cf: ComfyUI Custom node that supports face restore models and supports CodeFormer Fidelity parameter (github. Then i deleted and redownloaded comfyui and Match two faces' shape before using other face swap nodes Face-swapping tools typically only replace facial features during the swap, without altering the facial shape. it took a while to adjust the right setting in the face detailer, but im pretty happy with the outcome. Face recognition learning the faces? Both did not solved this, all is separated now and sd1. Both of my images have the flow embedded in the image so you can simply drag and drop the image into ComfyUI and it should open up the flow but I've also included the json in a zip file. Members Online • there are eye detection models for face-detailer. python setup. As I spoke about before: Understanding YOLO Models and which one to pick On the other hand :)) I also documented on my git that the hand fixing is not working always, especially when the picture is too zoomed out (hands too small), or when the hand is too much distorted. I did Install Missing Custom Nodes, Update All, and etc etc, but there are many issues every time I load the workflows, and it looks pretty complicated to solve it. Put ImageBatchToImageList > Face Detailer > ImageListToImageBatch > Video Combine. We test 2000+ images with ground truth annotations and calculate the mAP like COCO seriously, It it unusual it the Input "input_image" goes first now, it gives a correct bypass and also it is right to have the main input first; You can now save face models as "safetensors" files (ComfyUI\models\reactor\faces) and load them into ReActor implementing suuuuup, :Dso, with set latent noise mask, it is trying to turn that blue/white sky into a space ship, this may not be enough for it, a higher denoise value is more likely to work in this instance, also if you want to creatively inpaint then Stable Diffusion XL has trouble producing accurately proportioned faces when they are too small. Do I Need to Have Any Technical Knowledge to Work with ComfyUI ReActor Node? Not . Commit ff56485. It introduces the concept of face parsing with YOLO models to accurately detect facial features. inputs: faces; crop_size: size of the square cropped face image MediaPipe is a bit worse at detection, and can't run on GPU in Windows, though it's much faster on CPU compared to Insightface. I also tried moving them to a different folder in A1111 (models\embeddings), and changing names of the texuals (and using that name in the prompt). Total VRAM 24564 MB, total RAM 32581 MB Set vram state to: NORMAL_VRAM Device: cuda:0 NVIDIA GeForce RTX 4090 : cudaMallocAsync VAE dtype: torch. In this case, what might help is cranking up the hand Lora to 1. 15. I am on a windows 10 environment, working inside of the Pycharm terminal Face detection models. Face restore model: There are two models to choose from: CodeFormer (opens in a new tab) GFPGAN (opens in a new tab) After downloading the model, you need to place it in the ComfyUI\models\facerestore_models folder Also, if this is new and exciting to you, feel free to post, but don't spam all your work. Members Online. zip. Insightface is not automatically installed, if you wish to use it follow these instructions: If you have a working compile environment, installing it can be as easy as: pip install insightface You signed in with another tab or window. Included is face_yolov8m hand_yolov8s person_yolov8m deepfashion2_yolov8s They should offer better detection for their intended target but maybe take a little longer. The are colabs out there that do it too. ". This extension is Take a two-step method by utilizing two FaceDetailers to repair severely damaged faces. Also, if this is new and exciting to you, feel free to post, but don't spam all your work. Also the hand and face detection have never worked. You can also just search for "Face Restore" through the ComfyUI manager and you'll find it there as well. It's just magic animate that does that. The Simple Detector first performs primary detection using BBOX_DETECTOR and then generates SEGS using the intersected mask from the detection results of SEGM_DETECTOR or SAMDetector. 10 and 3. I have been playing with facial recognition a little bit in python, but have been having trouble with getting dlib to work. Members Online • I've been working on a rotation aware face detailer recently. This involves the process of mapping, aligning, and blending to make the swapped face look natural, and integrated into the source image. Not comfyui but facefusion has regional masking that maybe be helpful Options like skin, left-eyebrow','right-eyebrow', left-eye, right-eye', eye-glasses, nose I have not test it in comfyui, I test it offline with multiple models, including xl-base, counterfeit, blue pencil and so on. If just left the last line, still got errors. The FaceDetailer node is a combination of a hi there, I'm having the same issue, I have checked your link but it seems unrelated to my problem. Reply reply I updated comfyui, all is working now ! Face detection: Method of detecting faces, you can choose according to your actual situation, but I usually choose the first one. upvotes insightface: Use the Load InsightFacenode from ComfyUI_IPAdapter_plus; image; threshold: minimal confidence score for detection; min_size: minimum face size for detection; max_size: maximum face size for detection; outputs: faces; FaceDetails. this is impressive MediaPipe is a bit worse at detection, and can't run on GPU in Windows, though it's much faster on CPU compared to Insightface. It combines several AI models to handle the complex task of swapping faces: Face Detection: Uses the "retinaface_resnet50" model to find faces in He gives you links to the Python 3. sample face loader not working . The face restoration model only works with cropped face images. By using comfyui_facetools, you can solve common problems such as misaligned faces, inconsistent face detection, and difficulties in face inpainting. I've hooked everything up but there are three nodes left unconnected on the input. Plan and track work Discussions. If you wear a face mask too high or too close to your eyes, your iPhone might pip3 install face_recognition; face_recognition and dlib are not fully installed in specific , but default anaconda env. All features ltdrdata / ComfyUI-extension-tutorials Public. bfloat16 Using pytorch cross Extension for ComfyUI to evaluate the similarity between two faces - cubiq/ComfyUI_FaceAnalysis Plan and track work Code Review. The FaceDetailer node is a combination of a Detector node for face detection and a Ultralytics w/ detection model for eyes Send bbox to BBOX Combined node otherwise it will work on each eye individually Connect SAM so it only works on the eyes as opposed to the entire bbox Send the SEGS to Detailer node (Going off memory so sorry if something is slightly off) Anything relevant to living or working in Japan such as lifestyle, food, style, environment, education, technology, housing, work, immigration, sport etc. A face detection model is used to send a crop of each face found to the face restoration model. Members Online • DWPose: Using yolox_m. It seems like you are experiencing a multi-batch issue. 09, 3. 2 -The face detection is simply not very good. Ensure that you have write permissions for the specified directory. The FaceDetailer node is a combination of a Detector node for face detection and a Detailer node for image enhancement. The portraits generated are not even close. upvotes DZ FaceDetailer is a custom node for the "ComfyUI" framework inspired by !After Detailer extension from auto1111, it allows you to detect faces using Mediapipe and YOLOv8n to create masks for the detected faces. Am I missing something? ComfyUI seems to have downloaded some models for face/hand detection on using this node for the first time, but I'm not seeing their output. I changed it and cleared pycache but still no face detection. 5 model, i have no clue what is going on, i dont want to use sdxl cause its not great with details like some trained 1. device('cpu') OR. ) This project is designed to detect whether images generated by ComfyUI are Not Safe For Work (NSFW). ) A custom node for ComfyUI that allows you to perform lip-syncing on videos using the Wav2Lip model. format(len still not working even I've installed cmake and pip3 install face_recognition – Pyae Phyoe FaceDetailer in comfyUI not working, generates black square Question - Help hi, I'm been experimenting and trying to migrate my workflow from AUTO1111 to Comfy, and now I'm stuck while trying to reproduce the ADetailer step I use in AUTO1111 to fix faces; I'm using the ImpactPack's FaceDetailer node, but no matter what, it doesn't fix the face Face detection: Method of detecting faces, you can choose according to your actual situation, but I usually choose the first one. Does't work with you need a load image batch from dir (inspire) node, attach the image from that to a save face model node give the load image the directory path of your faces give the save face model a face_model_name queue the prompt then using a load Detect faces and focus on one of them. face module. I've reinstalled pip and face detailer is working again. Recently, BOOLEAN was added to ComfyUI and Impact Pack is updated to use it. After downloading the model, you need to place it in the ComfyUI\models\facerestore_models folder. (The most common case is when parts of folders are replaced with symbolic links. import numpy as np #for np. How comfyui_facetools Works Plan and track work Code Review. models exception), download this EVA-CLIP model manually, put the file to your ComfyUI/models/clipand restart ComfyUI. Please share your tips, tricks, and workflows for using this software to create your AI art. then inpaint with FaceIDv2 again with enough denoising to add back in details but not enough to change the Welcome to the unofficial ComfyUI subreddit. It consistently fails to detect a face if the mouth is wide open, for instance, or any kind of contorted facial features (even with the threshold set at the minimum) -If it doesn't detect a face, it stops the entire ComfyUI workflow. How to Compare your settings with the Face Detailer function of my AP Workflow for ComfyUI. Thank for Ur replay,Igot the face swap working but the face enhance didn't appear in Welcome to the unofficial ComfyUI subreddit. If this is your first dive into comfyUI, it can look intimidating with all the nodes and noodles. A lot of the options for Simswap are not exposed and the default values are being used. Face Predictor 81 landmarks and the Face Recognition models and place them into the dlib directory. - Welcome to the unofficial ComfyUI subreddit. face_locations(image) print("I found {} face(s) in this photograph. Instead of using techniques like virtual DOM diffing, Svelte writes code that surgically updates the DOM when the state of your app changes. The developer of the controlnet_aux preprocessors acknowledged that there is a bug. 5 has its own clip neg and positive that go to the pipe, still wont upscale the face wth sd1. This custom node enables you to generate new faces, replace faces, and perform other Welcome to the unofficial ComfyUI subreddit. Open Cassady8912 opened this issue Oct 3 Welcome to the unofficial ComfyUI subreddit. StyleHEAT: One-Shot one is the face (the redhead woman) and the other one is only the head position (the one from tron evolution). If there is more than one face, you need to make sure your index changes (in case you wanted to change the face of the 3rd face in the pic). This node is particularly useful for AI artists who want to improve the quality of faces in their artwork, whether they are working with realistic portraits or anime-style images. The reshaped face can be used as input for these tools, enhancing their performance and the quality of the final output. You switched accounts on another tab or window. It is less of a model and more like a "face preset". Leveraging advanced algorithms, DeepFuze enables users to combine audio and video with unparalleled realism, ensuring perfectly synchronized facial movements. Insightface is not automatically installed, if you wish to use it follow these instructions: If you have a working compile environment, installing it can be as easy as: pip install insightface Welcome to the unofficial ComfyUI subreddit. 3. Face Detection Failures: Make sure the input images are clear and Hey this is my first ComfyUI workflow hope you enjoy it! I've never shared a flow before so if it has problems please let me know. Execute install/(de)activation script for 'G:\AI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_FaceAnalysis' Collecting insightface Using cached insightface-0. What can I do? You don't necessarily need a PC to be a member of the PCMR. Yeah I’ve seen if I use the hand / face ones they work fine. With this configuration, you can easily configure different common inputs and streamline workflow modification using In this article, I will introduce you to Face Detailer, a collection of tools and techniques designed to fix faces and facial features. I was able to get it working, Bypass the AnimateDiff loader when inputting in the model in Facedetailer the See this image, Put ImageBatchToImageList > Face Detailer > ImageListToImageBatch but the images are not consistent : BTW I export the frames and fix the face in Adetailer in Automatic1111, Updating the current repo to include a ComfyUI custom node which allows aplying a face detection model - gopeto22/ComfyUI-Bringing-Old-Photos-Back-to-Life. c:\programdata\anaconda3\scripts\face_detection. Code; Issues 18; Pull requests 0; Actions; Projects 0; For demanding projects that require top-notch results, this workflow is your go-to option. Hi, i am trying to perform face swap on animal characters on children's storybooks. You signed out in another tab or window. It uses a machine learning model to classify images as either safe or not safe for work. It doesn't do head or person swap. Please keep posted images SFW. 0 for ComfyUI (Hand Detailer, Face Detailer, Free Lunch, Image Chooser, XY Plot, ControlNet/Control-LoRAs, Fine-tuned SDXL models, SDXL Base+Refiner, ReVision, Upscalers, Prompt Builder, Debug, etc. Densepose control net in comfyui does not do that when applied. WAW. I just went back to 4c81504 and now it seems to work. Nowhere in the range does the generated image follow this pose. Share. 5, 1. not sure what they're callrd Reply reply DeepFuze is a state-of-the-art deep learning tool that seamlessly integrates with ComfyUI to revolutionize facial transformations, lipsyncing, video generation, voice cloning, face swapping, and lipsync translation. load_image_file("My_Image. 7 _** ComfyUI start up time: 2023-10-20 23:13:01. ndarray type annotation class FaceAnalysis2(FaceAnalysis): # NOTE: allows setting Does't work with my installation. The FaceDetailer node is a combination of a Detector node for face detection and a Face detection: Method of detecting faces, you can choose according to your actual situation, but I usually choose the first one. The video demonstrates using ComfyUI's custom nodes for detailed face parsing and Welcome to the unofficial ComfyUI subreddit. In order for this to become a tool it needs to me functional not me working to get it functional. Follow Insightface is not automatically installed, if you wish to use it follow these instructions: If you have a working compile environment, installing it can be as easy as: pip install insightface or for the portable version, in the ComfyUI_windows_portable -folder: Welcome to the unofficial ComfyUI subreddit. then a face detailer makes the image close to the character (but not really perfect), a following low denoise face detailer does the final step. sam_model_opt and segm_detector_opt are both optional inputs, and if neither of them is provided, the detection results from BBOX are returned as is. Prestartup times for custom nodes: 0. detect_poses(detected_map,input_image, include_hand, include_face). When both inputs are Upload images, audio, and videos by dragging in the text input, pasting, or clicking here. The starting index of the detected faces list. This library uses dlib's implementation of face recognition algorithms, which is more accurate and faster than the algorithms included in the cv2. All features Documentation GitHub Skills Blog Solutions By size. It combines advanced face swapping and generation techniques to deliver high-quality outcomes, ensuring a comprehensive solution for your needs. By default, Face ID doesn't work if anything is covering your mouth and nose. 1' and I'm using cuda 12. 809618 Total VRAM 12288 MB, total RAM 32706 MB Set vram state to: NORMAL_VRAM Device: cuda:0 NVIDIA GeForce RTX 3060 : cudaMallocAsync VAE dtype: torch. device('privateuseone') This repository contains custom nodes designed to work with ComfyUI for face detection, restoration, and visualization using YuNet and dlib models. 11: Gourieff/comfyui-reactor-node: Fast and Simple Face Swap Extension Node for ComfyUI To get the just released IP-Adapter-FaceID working with ComfyUI IPAdapter plus you need to Face Detailer - Custom Node for ComfyUI (Stable Diffusion) Face Detailer is a custom node for the "ComfyUI" framework inspired by !After Detailer extension from auto1111, it allows you to detect faces using Mediapipe and YOLOv8n to create masks for the detected faces. It’s when I use the ones that have tags Face Fixer OpenCV (Mikey): FaceFixerOpenCV is a powerful node designed to detect and enhance faces within images using OpenCV's robust face detection algorithms. Finally I deleted all those lines and just used EITHER OF THESE. I can't for the life of me find how or where to get this input. The TLDR This tutorial explores advanced fine-tuning techniques for face detection and segmentation in AI-generated images, particularly for character faces in animations using Stable Diffusion. import face_recognition image = face_recognition. When there is a significant difference in facial shape between the target person and the person in the original photo, the result of the face swap is less satisfactory. onnx for bbox detection and dw-ll_ucoco_384. There are several models available to perform face restorations, as well as many interfaces; here I will focus on two solutions using ComfyUI and Stable-Diffusion-WebUI. 1 and Simswap. Next generation face toolkit for ComfyUI. I just put this together so that I could use Simswap in my workflows. Enterprise Saved searches Use saved searches to filter your results more quickly Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community. And Also Bypass the AnimateDiff Loader model to Original Model loader in the To Basic Pipe node else It will give you Noise on the face (as AnimateDiff loader dont work on single image, you need 4 atleast maybe and facedetailer can handle only 1 ) Only Drawback is there will be no This is the first time I see Face Hand adetailer in Comfyui workflow Also, if this is new and exciting to you, feel free to post, but don't spam all your work. It does the whole face, which fixes the eyes but also changes the other features. Have you tried changing the face detection model (I use YOLOv5l). There may be compatibility issues in future upgrades. A hacky implementation of Simswap based on Comfyui ReActor Node 0. dat shape_predictor_5_face_landmarks. Any time i try to get reactor working, i first in comfyui manager go to custom nodes and find reactor and install it. Today, I learn to use the FaceDetailer and Detailer (SEGS) nodes in the ComfyUI-Impact-Pack to fix small, ugly faces. put dlib model files in ComfyUI/models/dlib/ shape_predictor_68_face_landmarks. See the Advanced Tutorial for a more detailed explanation. This workflow is intended to be freely used and I do not mind if you want to share/post it elsewhere. I love the face detailer node - it does an amazing job. I have also tried getting it from insider custom nodes folder git clone command and after i get the plugin, I Since I used ComfyUI, I downloaded tons of workflows, but only around 10% of them work. If you want to make denseposes you can install detectron2 on Linux or WSL2 until he fixes it. Manage code changes Discussions. Collaborate outside of code Code Search. Support for various face detection models; Audio path upload for input audio file; Inputs. Notifications You must be signed in to change notification settings; [Feature]: Changing face detection model #445. The ReActor Node is a special tool in ComfyUI designed for face swapping. Solves 95% of faces too close not being detected for me. 7. It uses the face position and angle from the darker image and draws in that location the redhead woman. Collaborate outside of code Explore. The goal is to allow smoother implementation of face detection models. The face that has been damaged due to low resolution is restored with high resolution by generating and synthesizing it, in order to restore the details. 20K subscribers in the comfyui community. Typically, this phenomenon occurs when the user has modified some of the ComfyUI files. hi all, I've been using comfyui for about 10 days. You just have to love PCs. Now as pointers I will use a source video that is focus on the face or prepare a video that crops just the face portion and clean up the audio to be as perfect as possible, clear vocals no bg noise etc only work on small fragments of about 10 to 20 seconds of video at the time, and also use a image that is about the same size and frame face The "PreviewBridge" feature may not function correctly on ComfyUI versions released before July 1, 2023. Notifications You must be signed in to change notification settings; Fork 33; Star 593. AnimateDiff LCM workflow for separate foreground and background 0:07. py install --yes USE_AVX_INSTRUCTIONS --clean. NOTE: If you were permbanned for being nonresident prior to June of this year AND you have since moved to Japan, please contact the mod team to have your ban removed. py in src/controlnet_aux/dwpose it needs to have the include_hand and include_face bool def detect_poses(self, oriImg, include_hand=False, include_face=False) -> List[PoseResult]: and poses = self. Collaborate outside of code Face Predictor 5 landmarks, Face Predictor 81 landmarks and the Face Recognition models and place them into the dlib directory. It only works with ReActor and maybe other nodes using the same technology. (ComfyUI/models/dlib/). Ensure that the captured images have sufficient resolution and clarity for face detection. 0 seconds: F:\AI 2024\ComfyUI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes Expedition not working upvotes r/StableDiffusionInfo. Is basically like buying a creality 3d printer, instead of printing what you want you are tinkering with Whereas traditional frameworks like React and Vue do the bulk of their work in the browser, Svelte shifts that work into a compile step that happens when you build your app. I've been working on a rotation aware face detailer Hi i am also tring to solve roop quality issues,i have few fixes though right now I see 3 issues with roop 1 the faceupscaler takes 4x the time of faceswap on video frames 2 if there is lot of motion if the video the face gets warped with upscale 3 To process large number of videos pr photos standalone roop is better and scales to higher quality images but misses out on the img2img You signed in with another tab or window. If you're wearing a face mask and have already set up Face ID with a mask,* make sure that your eyes aren't blocked. Face detection/face swap for animal characters . 5. What we're trying to do is partial face replacement, maybe with masking out sections of the "load image" or something in conjunction with the ReActor but can't seem to figure it out. I tried to change the strength in the "Apply ControlNet (Advanced)" node from 0. Performs detailed work on SEGS without pasting it back onto the original image. If for some reason the auto-download fails (and you get face_analysis. py, init assert 'detection' in self. It is meant to be a faster solution to do "face" swap. number_of_faces: How many faces would you like to detect in total? (default: 5, min: 1, max: 100) start_index: Which face would you like to start with? (default: 0, step: 1). Reload to refresh your session. Speaking exercises not working on Android on 1 phone upvote I've been working on a rotation aware face detailer recently. 5 to 3. It takes an input video and an audio file and generates a lip-synced output video. When working with a large number of SEGS, such as Make Tile SEGS, it allows for monitoring the situation as improvements progress incrementally. Face detection models. If an image is classified as NSFW, an alternative image is returned. Follow the table of The "PreviewBridge" feature may not function correctly on ComfyUI versions released before July 1, 2023. Apply IPÄdapter FaceID + Load Insightface node not Nope, didn't work reinstalling it. upvotes ComfyUI ReActor Node replaces the faces in the source image with another face using advanced face detection technology. Images Not Saving: Check the path specified in the Save Image to Path node. Thank you and have a Great day :) It's a bit late, but the issue was that you needed to update your ComfyUI version to the latest one. These Models are the larger versions to face_yolov8s, hand_yolov8n and person_yolov8s. the first face detailer kind of primes the latent and a closer to the character image is the result. If you encounter any problems, please create an issue, thanks. GFPGAN. Members Online • Cropping Multi-face Detection Detected Faces and Saving as Image upvotes Expected Behavior Can not load PuLID Flux Actual Behavior Check the model and files, no problem , Steps to Reproduce The issue persists even after reinstalling the software and the Models. This then ensures a correct cropping node extracts the face around the edges. These methods involve removing distortions, adjusting DZ FaceDetailer is a custom node for the "ComfyUI" framework inspired by !After Detailer extension from auto1111, it allows you to detect faces using Mediapipe and YOLOv8n to create I found the issue - in the init. https://github. Significant Features Basic Auto Face Detection and Refinement. How many faces are appearing in the source and input image? If it's just one, make sure the face indexes are 0. Plan and track work Code Review. The face model is not similar to a checkpoint or a LoRA. Discuss all things about StableDiffusion here. png") face_locations = face_recognition. Ok solved it . Improve this answer. dat shape_predictor_81_face_landmarks. dat. The "PreviewBridge" feature may not function correctly on ComfyUI versions released before July 1, 2023. 0 seconds: D:\qiuye\ComfyUI\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager. And above all, BE NICE. GianoBifronte. ComfyUI_FaceShaper is designed to work seamlessly with other face-swapping nodes and tools, such as InstantID and DZ_FaceDetailer. The ComfyUI-Impact-Pack adds many Custom Nodes to [ComfyUI] “to conveniently enhance images through Detector, Detailer, dont update from manager, it aint working often. ADMIN MOD AP Workflow 4. Included in the ComfyUI Impact Pack is a collection of bespoke nodes that streamline and simplifies picture enhancement with the help of advanced tools like the Detector, Detailer, Upscaler, and Pipe. I've tried to incorporate FaceDetailer but as far as I can tell it's not possible to do only a part of the face. It's probably bloated with lots of unnecessary code and probably does not work optimally. safetensors . This is NO place to show-off ai art unless it's a highly educational Most consistent non-face/finger detection I've found so far is yolov8 skin and deepfashion that can very accurately and consistently detect skin and clothing, but that's not anatomy obviously, which leads to convoluted workflows, multiple mask operations and an unreliable mess overall. Nothing seems to work! 😔 To ComfyUi they don't seem to exist. images: Input video frames Related Work. Saved searches Use saved searches to filter your results more quickly I downloaded reactor and it was working just fine then i must have downloaded something that was interfering with it because i uninstalled everything via manager and it still didnt work. DeepFuze is a state-of-the-art deep learning tool that seamlessly integrates with ComfyUI to revolutionize facial transformations, lipsyncing, video generation, voice cloning, face swapping, and lipsync translation. A lot of people are just discovering this technology, and want to show off what they created. This project is under development. device = torch. It splits face recognition into small and large, treating them separately. I've I'm trying to use the FaceDetailer node from the ComfyUI Impact Pack. It's not about the hardware in your rig, but the software in your heart! Join us in celebrating and promoting tech, knowledge, and the best gaming, study, and work platform there exists. Face restore model: There are two models to choose from: CodeFormer. I'm having the same issue. for those with only one character, i Welcome to the unofficial ComfyUI subreddit. so you should to copy from. onnx for pose estimation Instead of using the cv2. extension? Gourieff / comfyui-reactor-node Public. Make use of FaceDetailer for simple, automated face improvement and detection. my onnxruntime is '1. But closer to 3, the image gets corrupted. Find more, search less Explore. These will automaticly be downloaded and placed in models/facedetection the first time each is used. anvdu ncxzxu fbgjor cagskrc ptytyfy ldsnmx nlzvpio ldhta pyrm gwjpq