Let's talk about the design concept of FCOS structure？

[TAOSHss]

❔Question

Let's talk about the design concept of FCOS structure in YOLOV5

Additional context

[github-actions[bot]]

👋 Hello @TAOSHss, thank you for your interest in 🚀 YOLOv5! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available.

For business inquiries or professional support requests please visit https://www.ultralytics.com or email Glenn Jocher at glenn.jocher@ultralytics.com.

Requirements

Python 3.8 or later with all requirements.txt dependencies installed, including torch>=1.7. To install run:

$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab and Kaggle 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 YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), testing (test.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

@TAOSHss hi there! FCOS is a very interesting paper, I especially liked the left-right-top-bottom interpretation of the regression space.

It would be an interesting experiment replacing the existing xywh boxes with lrtb boxes.

[TAOSHss]

What I said is that in the YOLOV5 network structure, the initial input part is called FCOS, but the function of this structure is to slice a large image into 4 small size and thin inputs

---Original--- From: "Glenn Jocher"<notifications@github.com> Date: Sat, Feb 27, 2021 03:08 AM To: "ultralytics/yolov5"<yolov5@noreply.github.com>; Cc: "TAOSHss"<764152567@qq.com>;"Mention"<mention@noreply.github.com>; Subject: Re: [ultralytics/yolov5] Let's talk about the design concept of FCOS structure？ (#2306)

@TAOSHss hi there! FCOS is a very interesting paper, I especially liked the left-right-top-bottom interpretation of the regression space.

It would be an interesting experiment replacing the existing xywh boxes with lrtb boxes.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[TAOSHss]

Focus not FCOS，im so sorry

---Original--- From: "Glenn Jocher"<notifications@github.com> Date: Sat, Feb 27, 2021 03:08 AM To: "ultralytics/yolov5"<yolov5@noreply.github.com>; Cc: "TAOSHss"<764152567@qq.com>;"Mention"<mention@noreply.github.com>; Subject: Re: [ultralytics/yolov5] Let's talk about the design concept of FCOS structure？ (#2306)

@TAOSHss hi there! FCOS is a very interesting paper, I especially liked the left-right-top-bottom interpretation of the regression space.

It would be an interesting experiment replacing the existing xywh boxes with lrtb boxes.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

[glenn-jocher]

@TAOSHss Oh, yes FCOS is a separate architecture that was quite performant: https://arxiv.org/abs/1904.01355

Focus() module in YOLOv5 is a space-to-channel conversion that helps speed up the the first convolution operations, which are otherwise quite slow due to the very large grids in the first convolution (i.e. native resolution image).

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.