Skip to content

Releases: henryzhuhr/deep-object-detect-track

v1.1.0 (2024.6.7)

07 Jun 08:24
Compare
Choose a tag to compare

Major Features and Improvements Summary

New Features:

  • Added semi-automatic annotation assistant tool

📎 Attachment:

  • Original yolov5s exported model
  • A sample dataset available for download
  • Model trained using the sample dataset and its exported model

主要功能和改进摘要

新功能:

  • 新增半自动化标注辅助工具

📎 附件:

  • 原始 yolov5s 导出模型
  • 一个样本数据集可供下载
  • 使用样本数据集训练的模型及其导出模型

v1.0.0 (2024.6.3)

03 Jun 07:03
Compare
Choose a tag to compare

Major Features and Improvements Summary

Features:

  • Automatically create and activate python environment for training and inference.
  • Object detection (Yolov5) and tracking(BYTETrack) algorithms are supported.
  • Tutorials for preparing dataset.
  • Train model with custom dataset.

Deployment support:

  • ONNXRuntime: inference with ONNXruntime are supported.
  • Openvino: inference with Openvino 2024.1.0 is supported, the previous version has not been rigorously tested.
  • TensorRT: inference with TensorRT 10.0.1 and 8.6.1 are supported.

📎 Attachment:

  • A sample dataset is prepared for download.

主要功能和改进摘要

功能:

  • 自动创建和激活 python 环境进行训练和推理。
  • 支持目标检测(Yolov5)和跟踪(BYTETrack)算法。
  • 准备数据集的教程。
  • 训练自己的数据集

部署支持:

  • ONNXRuntime:支持 ONNXruntime 的推理。
  • Openvino:支持使用 Openvino 2024.1.0 进行推理,对于先前的版本没有经过严格测试,
  • TensorRT:支持 TensorRT 10.0.1 和 8.6.1 的推理。

📎 附件:

  • 准备了一个示例数据集以供下载。