MT3

KO | EN

gitlite — search

#python #ai #llm #golang #go #android #machine-learning #java #windows #kubernetes #awesome-list #pytorch #security #open-source

(★ 48)

Code for the paper "Next Generation Multitarget Trackers: Random Finite Set Methods vs Transformer-based Deep Learning" (https://arxiv.org/abs/2104.00734).

MT3 최신버젼 다운로드

최종 버전 다운로드 (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

## Introduction
This repository contains the code for the paper "Next Generation Multitarget Trackers: Random Finite Set Methods vs Transformer-based Deep Learning" (https://arxiv.org/abs/2104.00734). The code for MT3 was developed as joint effort by Juliano Pinto, Georg Hess, William Ljungbergh, Yuxuan Xia, Lennart Svensson and Henk Wymeersch, and was partially based on the code available at the repositories for [DETR](https://github.com/facebookresearch/detr) and [Deformable-DETR](https://github.com/fundamentalvision/Deformable-DETR).

## Setting up
In order to set up a conda environment with all the necessary dependencies, run the command:
  ```
   conda env create -f conda-env/environment-<gpu/cpu>.yml
  ```

## Running experiments

Run an experiment using the `train.py` script. Example usage:

```
src/training.py -tp configs/tasks/task1.yaml -mp configs/models/mt3.yaml
```

Training hyperparameters such as batch size, learning rate, checkpoint interval, etc, are found in the file `configs/models/mt3.yaml`.

## Evaluating experiments

After an experiment has generated checkpoints, you can evaluate its average GOSPA score using the `eval.py` script. The evaluation hyperparameters can be found inside `configs/eval/default.yaml`. Example usage:

```
src/eval.py -rp src/results/experiment_name -tp configs/tasks/task1.yaml
```

MT3

MT3 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

MT3

MT3 최신버젼 다운로드

오늘의 추천 프로젝트

Was this content helpful?

Recent Feedback

Download README