Sam comfyui. You switched accounts on another tab or window.

Sam comfyui co/spaces/SkalskiP/florence-sam - ComfyUI When trying to select a mask by using "Open in SAM Detector", the selected mask is warped and the wrong size - before saving to the node. This is a ComfyUI node based-on Semantic-SAM official implementation. This node leverages advanced machine learning techniques to predict and create masks based on ControlNetApply (SEGS) - To apply ControlNet in SEGS, you need to use the Preprocessor Provider node from the Inspire Pack to utilize this node. It is not difficult to understand. Thanks! I dove in and have been messing with Comfyui now for the past couple of days, first with this, and then got into setting up controlnet and openpose. It has 7 workflows, including Yolo World ins You signed in with another tab or window. Willkommen zu einem neuen Video, in dem ich wieder Wissen mit Lebenszeit tausche. Segment Anything Model (SAM) arXiv: ComfyUI This tutorial focuses on Yolo World segmentation and advanced inpainting and outpainting techniques in Comfy UI. _rebuild_tensor_v2", "torch. How to use. Write prompt for naked body (very important, determines gender). This model is responsible for generating the embeddings from the input image. Based on GroundingDino and SAM, use semantic strings to segment any element in an image. pickle. First, ensure that Git is installed on your computer and that you installed ComfyUI . py. (Problem solved) I am a beginner at learning comfyui. Navigation Menu Toggle navigation. download models from Project. Three methods to remove background in ComfyUI. This will respect the nodes input seed to yield reproducible results like NSP and Wildcards. g. Currently, Impact Pack is providing the more sophisticated SAM model instead of the SEGM_MODEL for silhouette extraction. 8c62972 over 1 year ago. All kinds of masks will generate to choose. Safetensors. Compared with SAM, Semantic-SAM has better fine-grained capabilities and more candidate masks. , Karras). Look at blue boxes from left to right, and choose the best mask at every stage by Use the sam_vit_b_01ec64. ***************************************************It seems there is an issue with gradio. Empowers AI Art creation with high-speed GPUs & efficient workflows, no tech setup needed. segs_preprocessor and control_image can be selectively applied. Detected Pickle imports (3) "torch Contribute to ltdrdata/ComfyUI-extension-tutorials development by creating an account on GitHub. Enter ComfyUI SAM2(Segment Anything 2) in the search bar After installation, click the Restart button to restart ComfyUI. This version is much more precise and There are multiple options you can choose with: Base, Tiny, Small, Large. com/ltdrdata/ComfyUI SAM Overview. Click on an object in the first view of source views; SAM segments the object out (with three possible masks);; Select one mask; A tracking model such as OSTrack is ultilized to track the object in these views;; SAM segments the object out in each Comfy Summit Workflows (Los Angeles, US & Shenzhen, China) Challenges. It looks like the whole image is offset. dirname(os. py", line 317, in execute output_data, output_ui, has_subgraph = get_output Welcome to the unofficial ComfyUI subreddit. SAM is designed to Based on GroundingDino and SAM, use semantic strings to segment any element in an image. pth" model - download (if you don't have it) and put it into the "ComfyUI\models\sams" directory; Use this Node to gain the best results of the face swapping process: ReActorImageDublicator Node - rather useful for those who create videos, it helps to duplicate one image to several frames to use them with VAE Download sam_vit_h，sam_vit_l, sam_vit_b, sam_hq_vit_h, sam_hq_vit_l, sam_hq_vit_b, mobile_sam to ComfyUI/models/sams folder. history blame No virus pickle. ComfyUI nodes to use segment-anything-2. _rebuild From this menu, you can either open a dialog to create a SAM Mask using 'Open in SAM Detector', or copy the content (likely mask data) using 'Copy (Clipspace)' and generate a mask using 'Impact SAM Detector' from the clipspace menu, and then paste it It seems your SAM file isn't valid. com/workflows/b68725e6-2a3d-431b-a7d3-c6232778387d https://github. If necessary, you can find and redraw people, faces, and hands, or perform functions such as resize, resample, and add noise. py at main · storyicon/comfyui_segment_anything Based on GroundingDino and SAM, use semantic strings to segment any element in an image. FloatStorage" What is From this menu, you can either open a dialog to create a SAM Mask using 'Open in SAM Detector', or copy the content (likely mask data) using 'Copy (Clipspace)' and generate a mask using 'Impact SAM Detector' from the clipspace menu, and then paste it Explore the newest features, models, and node updates in ComfyUI and how they can be applied to your digital creations. However, it is recommended to use the PreviewBridge and Open in SAM Detector approach instead. Node options: sam_model: Select the SAM model. This is also the reason why there are a lot of custom nodes in this workflow. FLUX Tools Inpaint Basic with SAM. FloatStorage", "collections. Run it. bb894b1 verified 1 day ago. This version is much more precise and ComfyUI custom node implementing Florence 2 + Segment Anything Model 2, based on SkalskiP's space at https://huggingface. - comfyui_segment_anything/node. You signed in with another tab or window. Contribute to ycyy/ComfyUI-Yolo-World-EfficientSAM development by creating an account on GitHub. Reply reply Contribute to Bin-sam/DynamicPose-ComfyUI development by creating an account on GitHub. Usage: ComfyUI nodes to use segment-anything-2. com/continue-revolution/sd-webui-segment-a I have ensured consistency with sd-webui-segment-anything in terms of output when given the same input. We provide a workflow node for one-click segment. And it doesn't StableSAM / sam_vit_h_4b8939. ; If set to control_image, you can preview the cropped cnet image through If you're running on Linux, or non-admin account on windows you'll want to ensure /ComfyUI/custom_nodes and comfyui_controlnet_aux has write permissions. segment anything's webui. Select a model. Share and Run ComfyUI workflows in the cloud. Unlike MMDetDetectorProvider, for segm models, BBOX_DETECTOR is also provided. However, I found that there is no Open in MaskEditor button in my node. Whether you're working on Checkpoints of BrushNet can be downloaded from here. path: sys. I tried using sam: models\\sam under my a1111 section. As we wrap up ComfyUI Node that integrates SAM2 by Meta. This node have been valided on Ubuntu-20. Leveraging advanced algorithms, DeepFuze enables users to combine audio and video with unparalleled realism, ensuring perfectly synchronized facial movements. RunComfy: Premier cloud-based Comfyui for stable diffusion. md. Please ensure that you use SAMLoader (Impact) as instructed in the message. The quality and type of the embeddings depend on the specific SAM model used. You can composite two images or perform the Upscale The Segment Anything Model (SAM) produces high quality object masks from input prompts such as points or boxes, and it can be used to generate masks for all objects in an image. SAM is a detection feature that get segments based on specified position, and it doesn't have the capability to detect based on tags. history blame contribute delete 454 Bytes. Load picture. Your question 在最新版本comfyui上运行“segmentation”功能的节点在加载SAM模型时会出现这个报错。我分别尝试了“comfyui_segment You signed in with another tab or window. I have updated the requirements. Mastering Inpainting in ComfyUI with SAM (Segment Anything) Updated: 1 comfyui-extension-models. Detectors. The checkpoint in segmentation_mask_brushnet_ckpt provides checkpoints trained on BrushData, which has segmentation prior (mask are with the same shape of objects). In the second case, I tried the SAM Detector both in front of CLIPSeg and as an auxiliary model to the Simple Detector after CLIPSeg. Recently I want to try to detect certain parts of the image and then redraw it. 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. ComfyUI enthusiasts use the Face Detailer as an essential node. In this video, we show how you can easily and accurately mask objects in your video using Segment Anything 2 or SAM 2. If you have another Stable Diffusion UI you might be able to reuse the dependencies. ComfyUI_LayerStyle / ComfyUI / models / EVF-SAM / evf-sam / README. If you have installed ComfyUI-Manager, you can update comfyui with it. insert(0, current Welcome to the unofficial ComfyUI subreddit. Please keep posted images SFW. EVF-SAM extends SAM's capabilities with text-prompted segmentation, achieving high accuracy in Referring Expression Segmentation. License: openrail. EVF-SAM EVF-SAM: Early Vision-Language Fusion for Text-Prompted Segment Anything Model. bf831f0 verified 8 months ago. bf831f0 verified 11 months ago. history blame contribute delete Safe. _utils. SAM Parameters: Define your SAM parameters for segmentation of a image; SAM Parameters Combine: Combine SAM parameters; SAM Image Mask: SAM image masking You signed in with another tab or window. Contribute to neverbiasu/ComfyUI-SAM2 development by creating an account on GitHub. But I still think the result turned out pretty well and wanted to share it with the community :) It's pretty self-explanatory. A ComfyUI extension for Segment-Anything 2. There is a good comparison between the three tested workflows for face detailers, and you can decide which workflow you prefer. If a control_image is given, segs_preprocessor will be ignored. Extensions; segment anything; ComfyUI Extension 791 stars. Created by: Can Tuncok: This ComfyUI workflow is designed for efficient and intuitive image manipulation using advanced AI models. - comfyanonymous/ComfyUI Share and Run ComfyUI workflows in the cloud. metadata. # ComfyUI SAM2 (Segment Anything 2) This project adapts the SAM2 to incorporate functionalities from [comfyui_segment_anything] Ready to take your image editing skills to the next level? Join me in this journey as we uncover the most mind-blowing inpainting techniques you won't believ In this blog post, we will delve into the implementation of SAM 2 within the ComfyUI environment, a powerful and user-friendly platform for exploring and leveraging the capabilities By using the segmentation feature of SAM, it is possible to automatically generate the optimal mask and apply it to areas other than the face. SAM2 is trained on real-world videos and masklets and can be applied to image alteration, In this video, we show how you can easily and accurately mask objects in your video using Segment Anything 2 or SAM 2. - antoinedelplace/comfyui SAM Overview. Segment Anything Model 2 (SAM 2) is a continuation of the Segment Anything project by Meta AI, designed to enhance the capabilities of automated image A ComfyUI custom node designed for advanced image background removal and object segmentation, utilizing multiple models including RMBG-2. abspath(__file__)) # Add the current directory to the first position of sys. There is now a install. It seems that until there's an unload model node, you can't do this type of heavy lifting using multiple models in the same Created by: rosette zhao: (This template is used for Workflow Contest) What this workflow does 👉This workflow uses interactive sam to select any part you want to separate from the background (here I am selecting person). chflame163 Upload 7 files. segmentation_mask_brushnet_ckpt This image, plus your prompt and input settings are sent to the awesome ComfyUI workflow executor by fal. 35cec8d verified 29 days ago. Thanks ,I will check , and where can I find some same model that support hq? The GitHub repository “ComfyUI-YoloWorld-EfficientSAM” is an unofficial implementation of YOLO-World and EfficientSAM technologies for ComfyUI, aimed at enhancing object detection and ComfyUI-YoloWorld-EfficientSAM creating "tmp" folder in main directory of drive. I love all the flexibility available for ComfyUI, you ComfyUI_SemanticSAM. I'm using an SDXL Lightning checkpoint for fast inference. Authored by WASasquatch. Belittling their efforts will get you banned. abhishek HF staff add weights. (SAM 2) to offer promptable visual segmentation, making it easier to isolate and manipulate specific parts of an image or video. ComfyUI Segment Anything: Based on GroundingDino and SAM, use semantic strings to segment any element in an image. preview code | raw Copy download link. If it does not work, ins There is discussion on the ComfyUI github repo about a model unload node. fastsam mobilesam EfficientSAM. SAM Editor assists in generating silhouette masks usin Welcome to the unofficial ComfyUI subreddit. Users can take this node as the pre-node for inpainting to obtain the mask region. chflame163 Upload 14 files. As well as "sam_vit_b_01ec64. FloatStorage", "torch. Also, if this is new and exciting to you, feel free to ComfyUI-Workflow-Component provides functionality to simplify workflows by turning them into components, as well as an Image Refiner feature that allows improving images based on components. Toggle theme Login. The image on the left is the original image, the Ready to take your image editing skills to the next level? Join me in this journey as we uncover the most mind-blowing inpainting techniques you won't believ The SAM Image Mask node is designed to generate object masks from an input image using the Segment Anything Model (SAM). How to use this workflow You signed in with another tab or window. 0 but my laptop with a RTX 3050 Laptop 4GB vRAM was not able to generate in less than 3 minutes, so I spent some time to get a good configuration in ComfyUI, now I get can generate in 55s (batch images) - 70s (new prompt detected) getting a great images after the refiner kicks in. 04 This guide has taken us on an exploration of the art of inpainting using ComfyUI and SAM (Segment Anything) starting from the setup, to the completion of image rendering. pth as the SAM_Model. 57K ComfyUI SAM2(Segment Anything 2) This project adapts the SAM2 to incorporate functionalities from comfyui_segment_anything. This project adapts the SAM2 to incorporate functionalities from a/comfyui_segment_anything. With a single click on an object in the first view of source views, Remove Anything 3D can remove the object from the whole scene!. license: apache-2. riwa. pth. Please share your tips, tricks, and workflows for using this software to create your AI art. , Euler A) from the scheduler (e. Updating ComfyUI for Manual Git Installations. You switched accounts on another tab or window. This version is much more precise and practical than the first version. One thing to note is that ComfyUI separates the sampler (e. Detection method: GroundingDinoSAMSegment (segment anything) device: Mac arm64(mps) But in this process, for my example picture, if it is the head, it can be detected, but there is no accurate way to detect the arms, waist, chest, etc. 0. The Impact Pack's Detector includes three main types: BBOX, SEGM, and SAM. And above all, BE NICE. Follow the ComfyUI manual installation instructions for Windows and Linux. _rebuild_tensor_v2", "collections. It has been trained on a dataset of 11 million images and 1. The process begins with the SAM2 model, which allows for precise segmentation and masking of ComfyUI custom node implementing Florence 2 + Segment Anything Model 2, based on SkalskiP's HuggingFace space. Detected Pickle imports (3) "torch. UltralyticsDetectorProvider - Loads the Ultralystics model to provide SEGM_DETECTOR, BBOX_DETECTOR. Samplers determine how a latent is denoised, schedulers determine how much noise is removed per step. See all posts. -multimask checkpoints are jointly trained on Ref, ADE20k First and foremost, I want to express my gratitude to everyone who has contributed to these fantastic tools like ComfyUI and SAM_HQ. Traceback (most recent call last): File "K:\ComfyUI_windows_portable\ComfyUI\execution. You signed out in another tab or window. com/LykosAI/StabilityMatrixhttps://github. Then, manually refresh your browser to clear the cache and access the updated list of nodes. If using mask-area, only some of the Created by: CgTopTips: In this video, we show how you can easily and accurately mask objects in your video using Segment Anything 2 or SAM 2. Welcome to the unofficial ComfyUI subreddit. path if current_directory not in sys. Reload to refresh your session. . OrderedDict" What is a pickle import? Use HQ-SAM with Grounding DINO can do it by input prompts here is a blog link, the author provides a workflow that you can download. The abstract of the paper states: ComfyUI_LayerStyle / ComfyUI / models / sams / sam_vit_h_4b8939. In the mean time, in-between workflow runs, ComfyUI manager has a "unload models" button that frees up memory. ControlNetApply (SEGS) - To apply ControlNet in SEGS, you need to use the Preprocessor Provider node from the Inspire Pack to utilize this node. Contribute to kijai/ComfyUI-segment-anything-2 development by creating an account on GitHub. Workflow Templates Get SAM Embedding Input Parameters: sam_model. #98 opened Dec 2, 2024 by thrabi 路径不要有中文 Put it in “\ComfyUI\ComfyUI\models\sams\“. ground_dino_model: Select the Grounding DINO model. ICU. By combining the object recognition capabilities of Florence 2 with the precise segmentation prowess of SAM 2, we can achieve remarkable results in Welcome to the unofficial ComfyUI subreddit. 1 billion masks, and has strong zero-shot performance on a variety of segmentation tasks. Images contains workflows for ComfyUI. A lot of people are just discovering this technology, and want to show off what they created. Fast and Simple Face Swap Extension Node for ComfyUI. The ComfyUI-Impact-Pack adds many Custom Nodes to [ComfyUI] “to conveniently enhance images through Detector, Detailer, Welcome to the unofficial ComfyUI subreddit. I am not sure if I should install a custom node or fix settings. Published on March 3, 2024How to Remove Background in ComfyUI? sam (download encoder, download decoder, source): A pre-trained model for any use cases. ComfyFlow Explore Blog Docs Menu. Save the respective model inside "ComfyUI/models/sam2" folder. 13a2313 verified 9 months ago. py --force-fp16. EVF-SAM is designed for efficient computation, enabling rapid inference in few seconds per image on a T4 GPU. The random_mask_brushnet_ckpt provides a more general ckpt for random mask shape. Kijai is a very talented dev for the community and has graciously blessed us with an early release. Looking at the repository, the code we'd be interested in is located in grounded_sam_demo. ReActor / models / sams / sam_vit_b_01ec64. txt file. The second image is the screenshot of my ComfyUi that does not have Open in MaskEditor and Is it solved, I'm also experiencing this situation, and it doesn't work even if I uninstall yolo. Alternative: Navigate to ComfyUI Manager I used this as motivation to learn ComfyUI. Extensions; WAS Node Suite; ComfyUI Extension: WAS Node Suite. like 3. Explore Docs Pricing. Author Fannovel16 (Account age: 3127days) Extension ComfyUI's ControlNet Auxiliary Preprocessors Latest Updated 2024-06-18 Github Stars 1. ComfyUI provides a bit more flexibility with this approach but does mean you can use a sampler/scheduler combo that produces crap images Updating ComfyUI for users who have installed ComfyUI-Manager. I follow the video guide to right-click on the load image node. (I You signed in with another tab or window. The sam_model parameter expects an AV_SAM_MODEL type, which is a pre-trained Segment Anything Model. 0, INSPYRENET, BEN, SAM, and GroundingDINO. Special thanks to storyicon for their initial implementation, which inspired me to create this repository. Segment Anything Model 2 (SAM 2) is a continuation of the Segment Anything project by Meta AI, designed to enhance the capabilities of automated image ComfyUI Yolo World EfficientSAM custom node. SAM 2. The comfyui version of sd-webui-segment-anything. Use this model 5e06234 comfyui-extension-models / ComfyUI-Impact-Pack / sam_vit_b_01ec64. The default downloaded bbox model currently only detects the face area as a rectangle, and the segm model detects the SAM Parameters (SAM Parameters): Facilitates creation and manipulation of parameters for image segmentation and masking tasks in SAM model. SAMLoader - Loads the SAM model. Currently, since it's not merged, you can use this instead for immediate use: (my forked version) Hey guys, I was trying SDXL 1. I'm using an SDXL Lightning Welcome to the ComfyUI Group! This is the ultimate community for all things ComfyUI! Whether you're a developer, designer, hobbyist, or just someone passionate about ComfyUI, you're in the right The SAM (Segment Anything Model) node in ComfyUI integrates with the YoloWorld object detection model to enhance image segmentation tasks. Do not modify the file names. RdancerFlorence2SAM2GenerateMask - the node is self Together, Florence2 and SAM2 enhance ComfyUI's capabilities in image masking by offering precise control and flexibility over image detection and segmentation. Hope everyone By using PreviewBridge, you can perform clip space editing of images before any additional processing. ONNX. ; When setting the detection-hint as mask-points in SAMDetector, multiple mask fragments are provided as SAM prompts. Custom Nodes (5)GroundingDinoModelLoader (segment CLIPTextEncode (NSP) and CLIPTextEncode (BlenderNeko Advanced + NSP): Accept dynamic prompts in <option1|option2|option3> format. path. Here is an example of another generation using the same workflow. Many thanks to continue-revolution for their foundational work. The methods demonstrated in this aim to make intricate processes more accessible providing a way to express creativity and achieve accuracy in editing images. These are exceptionally well-crafted works, and I salute the creators. ¡Bienvenido al episodio 8 de nuestra serie de tutoriales sobre ComfyUI para para Stable Diffusion!En este video, te presentamos un tutorial completo sobre Co One of the key strengths of SAM 2 in ComfyUI is its seamless integration with other advanced tools and custom nodes, such as Florence 2, a vision-enabled large language model developed by Microsoft. Learn how to install and use SAM2, an open-source model for object segmentation, with ComfyUI, a custom node for Blender. Today, I learn to use the FaceDetailer and Detailer (SEGS) nodes in the ComfyUI-Impact-Pack to fix small, ugly faces. *Or download them from GroundingDino models on BaiduNetdisk and SAM models on BaiduNetdisk. Diffusers. SEGS is a comprehensive data format that includes information required for Detailer operations , such as masks , bbox , crop regions , confidence , label , and ComfyUI Node: SAM Segmentor Class Name SAMPreprocessor Category ControlNet Preprocessors/others. Launch ComfyUI by running python main. I'm not too familiar with this stuff, but it looks like it would need the grounded models (repo etc) and some wrappers made out of a few functions found in the file you linked (mask extraction nodes and for the main get_grounding_output method) Stable Diffusion XL has trouble producing accurately proportioned faces when they are too small. Do not use the SAMLoader provided by other custom nodes. And Impact's SAMLoader doesn't support hq model. Segment Anything Model (SAM): a new AI model from Meta AI that can "cut out" any object, in any image, with a single click SAM is a promptable segmentation system with zero-shot generalization to unfamiliar objects and images, without I'm trying to add my SAM models from A1111 to extra paths, but I can't get Comfy to find them. Sign in Yes, to use this, you'll need to install ComfyUI-YoloWorld-EfficientSAM. Write prompt for the whole picture (barely important). _rebuild_tensor_v2" BMAB is an custom nodes of ComfyUI and has the function of post-processing the generated image according to settings. and using ipadapter attention masking, you can assign different styles to the person and background by load different style pictures. That has not been implemented yet. Comfy. bat you can run to install to portable if detected. The Detector detects specific regions based on the model and returns processed data in the form of SEGS. Download the model files to models/sams under the ComfyUI root directory. Including: LayerMask: BiRefNetUltra, LayerMask: BiRefNetUltraV2, LayerMask: LoadBiRefNetModel, LayerMask: LoadBiRefNetModelV2, Created by: CgTopTips: In this video, we show how you can easily and accurately mask objects in your video using Segment Anything 2 or SAM 2. I haven't seen this, but it looks promising. ComfyUI-YOLO: Ultralytics-Powered Object Recognition for ComfyUI - kadirnar/ComfyUI-YOLO I am a newbie in ComfyUI. download Copy download link. Otherwise it will default to system and assume you followed ConfyUI's manual installation steps. It's not a straight forward as I was hoping to get good results, but tools like IPAdapter definitely move it in the right direction. ; The various models available in UltralyticsDetectorProvider can be downloaded through ComfyUI In order to prioritize the search for packages under ComfyUI-SAM, through # Get the absolute path of the directory where the current script is located current_directory = os. Summary. Note that --force-fp16 will only work if you installed the latest pytorch nightly. About Impact-Pack. Contribute to creeponsky/SAM-webui development by creating an account on GitHub. Install the ComfyUI dependencies. Updated: Dec 6, 2024 4:53 AM ComfyUI Segment Anything项目在ComfyUI框架下实现了核心功能，并提供详细的Python依赖安装指南和模型下载方式，确保与sd-webui-segment-anything一致。用户可以通过pip命令快速安装所需依赖，并自动或手动下载BERT、GroundingDino和SAM模型。如下载速度较慢，可设置代理加 Original sam is too slow, now there are some replacement, eg. Skip to content. Detected Pickle imports (3) "torch Split some nodes of the dependencies that are prone to problems into ComfyUI_LayerStyle_Advance repository. Check ComfyUI/models/sams. https://comfyworkflows. Create a "sam2" folder if not exist. - 1038lab/ComfyUI-RMBG The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface. licyk Upload 3 files. 125bdfb over 1 year ago. Gourieff Upload sam_vit_b_01ec64. This project is a ComfyUI version of https://github. The model can be used to predict segmentation masks of any object of interest given an input image. SAM (Segment Anything Model) was proposed in Segment Anything by Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alex Berg, Wan-Yen Lo, Piotr Dollar, Ross Girshick. Heute nehmen wir uns das faszinierende SAM-Modell vor - das Segment-Anythin Together, Florence2 and SAM2 enhance ComfyUI's capabilities in image masking by offering precise control and flexibility over image detection and segmentatio comfyui-extension-models / ComfyUI-Impact-Pack / sam_vit_b_01ec64. Click the ComfyUI Manager icon in the system tray, then click Update ComfyUI. Support. OrderedDict", "torch. ComfyUI-segment-anything-2 provides nodes for utilizing the segment-anything-2 tool by Facebook Research, enabling efficient image and video segmentation within the ComfyUI framework. I have the most up-to-date ComfyUI and ComfyUI-Impact-Pack change sam_vit_h to sam_vit_l to save memory. If a Mask Pointer is an approach to using small masks indicated by mask points in the detection_hint as prompts for SAM. eqlth cme wtaoyj kpqip dvzq wzxxtc cpsfhcl dsz wewuf fbd