embedding projector about ultralytics HOT 4 OPEN

👋 Hello @bonseong11, thank you for your interest in Ultralytics YOLOv8 🚀! We recommend a visit to the Docs for new users where you can find many Python and CLI usage examples and where many of the most common questions may already be answered.

If this is a 🐛 Bug Report, please provide a minimum reproducible example to help us debug it.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset image examples and training logs, and verify you are following our Tips for Best Training Results.

Join the vibrant Ultralytics Discord 🎧 community for real-time conversations and collaborations. This platform offers a perfect space to inquire, showcase your work, and connect with fellow Ultralytics users.

Install

Pip install the ultralytics package including all requirements in a Python>=3.8 environment with PyTorch>=1.8.

pip install ultralytics

Environments

YOLOv8 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all Ultralytics CI tests are currently passing. CI tests verify correct operation of all YOLOv8 Modes and Tasks on macOS, Windows, and Ubuntu every 24 hours and on every commit.

from ultralytics.

Hello! Thanks for reaching out with your question. 👋

Currently, YOLOv8 does not have a built-in feature exactly like TensorBoard's Embedding Projector. However, you can certainly use TensorBoard with YOLOv8 models for visualizing embeddings by extracting them separately during your model's evaluation stages and logging them to TensorBoard.

Here’s a quick code snippet on how you might log embeddings:

from torch.utils.tensorboard import SummaryWriter
import torch

# Example tensor of embeddings and labels
embeddings = torch.randn(100, 512)  # example embeddings
labels = torch.randint(0, 10, (100,))  # example labels

writer = SummaryWriter()
writer.add_embedding(embeddings, metadata=labels)
writer.close()

Make sure to align this with how you handle your YOLOv8 outputs! If you need more specific guidance on integrating this with YOLOv8 or any other tasks, feel free to ask. 😊

from ultralytics.

What should I do if I apply the embedding value to the learning data using the yolo(detect) best.pt model?

from ultralytics.

Hello! To apply the embedding values from a YOLO model trained for detection (like your best.pt model), you'd typically first extract the embeddings during or after inference, then use these embeddings accordingly in your learning data.

Here’s a concise example of how you might extract these embeddings from the model:

from ultralytics import YOLO
import torch

# Load your trained model
model = YOLO('path/to/best.pt')

# Assuming 'images' is your batch of input images
images = torch.rand((1, 3, 640, 640))  # dummy data, replace with actual image tensor

# Get embeddings
with torch.no_grad():
    model.eval()
    embeddings = model(images, embed=True)  # Set embed=True to get embeddings

# Now you can use 'embeddings' in your learning data

Make sure that your best.pt model is set up to return embeddings. If it isn't, you may need to modify the model definition slightly. If you need more specific help on this, don't hesitate to ask. Happy coding! 😊

from ultralytics.