Segment-Everything-Everywhere-All-At-Once
segment-anything
Segment-Everything-Everywhere-All-At-Once | segment-anything | |
---|---|---|
6 | 56 | |
4,064 | 44,293 | |
2.8% | 2.1% | |
7.9 | 0.0 | |
about 1 month ago | 28 days ago | |
Python | Jupyter Notebook | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
Segment-Everything-Everywhere-All-At-Once
-
Is supervised learning dead for computer vision?
Yes, you can. The model that I was talking about LLaVA only output text but other models such as SEEM (https://github.com/UX-Decoder/Segment-Everything-Everywhere-...) outputs a segmentation map. You could prompt the model "Where is the pickleball in the image?" and get a segmentation map that you could then use to compute its center. Please let me know if you would be interested to have SEEM available in Datasaurus
-
The less i know the better
I think people are just seeing the rate of progress and rightfully think that this stuff will be possible at some point. For the rotoscoping for example, here's an example of progress being made on that.
-
A robot showing off his moves
Yeah, it's definitely possible especially with all the recent advances. With segment anything systems (like SAM) and segmentation on NeRF reconstructions already being a thing the feasibility of this is more a time investment thing. Naive "scene understanding" is already possible in a few AR headsets at real-time, but the new papers in the past few weeks have made this much more trivial and faster to implement.
- Seem: Segment Everything Everywhere All at Once
-
[R] SEEM: Segment Everything Everywhere All at Once
Play with the demo on GitHub! https://github.com/UX-Decoder/Segment-Everything-Everywhere-All-At-Once
segment-anything
-
What things are happening in ML that we can't hear oer the din of LLMs?
- segment anything: https://github.com/facebookresearch/segment-anything
-
Zero-Shot Prediction Plugin for FiftyOne
In computer vision, this is known as zero-shot learning, or zero-shot prediction, because the goal is to generate predictions without explicitly being given any example predictions to learn from. With the advent of high quality multimodal models like CLIP and foundation models like Segment Anything, it is now possible to generate remarkably good zero-shot predictions for a variety of computer vision tasks, including:
-
Generate new version of a living-room with specific furniture
Render a new living room using a controlnet model of your choice to keep the basic structure. Load the original living room image and look for the furniture you want to change with a Segment Anything Model to create a mask. Use that mask on the new living room to inpaint new furniture.
-
How Do I read Github Pages? It is so exhausting, I always struggle, oh and I am on windows
Hello,So I am trying to run some programs, python scripts from this page: https://github.com/facebookresearch/segment-anything, and found myself spending hours without succeeding in even understanding what's is written on that page. And I think this is ultimately related to programming.
-
Autodistill: A new way to create CV models
Some of the foundation/base models include: * GroundedSAM (Segment Anything Model) * DETIC * GroundingDINO
-
How to Fine-Tune Foundation Models to Auto-Label Training Data
Webinar from last week on how to fine-tune VFMs, specifically Meta's Segment Anything Model (SAM).
What you'll need to follow along the fine-tuning walkthrough:
Images, ground-truth masks, and optionally, prompts from the Stamp Verification (StaVer) Dataset on Kaggle (https://www.kaggle.com/datasets/rtatman/stamp-verification-s...)
Download the model weights for SAM the official GitHub repo (https://github.com/facebookresearch/segment-anything)
Good understanding of the model architecture Segment Anything paper (https://ai.meta.com/research/publications/segment-anything/)
GPU infra the NVIDIA A100 should do for this fine-tuning.
Data curation and model evaluation tool Encord Active (https://github.com/encord-team/encord-active)
Colab walkthrough for fine-tuning: https://colab.research.google.com/github/encord-team/encord-...
I'd love to get your thoughts and feedback. Thank you.
-
Deploying a ML model (segment-anything) to GCP - how would you do it?
I now want users to be able to use the segment-anything model (https://github.com/facebookresearch/segment-anything) in my app. It's in pytorch if that matters. How it should work is that
-
The Mathematics of Training LLMs
Yeah, they are great and some of the reason (up the causal chain) for some of the work I've done! Seems really fun! <3 :))))
Facebook's Segment Anything Model I think has a lot of potentially really fun usecases. Plaintext description -> Network segmentation (https://github.com/facebookresearch/segment-anything/blob/ma...) Not sure if that's what you're looking for or not, but I love that impressing your kids is where your heart is. That kind of parenting makes me very, very, very, happy. :') <3
-
How hard is it to "code" a tool based on segment-anything and Stable diffusion ?
There are some snippets of Python code on the segment-anything github readme that show how to do this. Once you have it installed you can import functions from the segment-anything module, load a segmentation model, and generate masks for input images that match the prompt of your choice. You don't need Stable Diffusion for this, but you could load it through diffusers to do things like inpaint your images using the masks.
- The less i know the better
What are some alternatives?
Segment-Everything-Everywhere-
backgroundremover - Background Remover lets you Remove Background from images and video using AI with a simple command line interface that is free and open source.
LLaVA - [NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.
ComfyUI-extension-tutorials
guidance - A guidance language for controlling large language models.
stable-diffusion-webui-Layer-Divider - Layer-Divider, an extension for stable-diffusion-webui using the segment-anything model (SAM)
LoRA - Code for loralib, an implementation of "LoRA: Low-Rank Adaptation of Large Language Models"
Grounded-Segment-Anything - Grounded-SAM: Marrying Grounding-DINO with Segment Anything & Stable Diffusion & Recognize Anything - Automatically Detect , Segment and Generate Anything
autodistill - Images to inference with no labeling (use foundation models to train supervised models).
GroundingDINO - Official implementation of the paper "Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection"
CLIP - CLIP (Contrastive Language-Image Pretraining), Predict the most relevant text snippet given an image
mmdetection - OpenMMLab Detection Toolbox and Benchmark