Skip to content

Microsoft VoTT

Jump to bottom
Giovanni Emanuele Nocco edited this page Mar 11, 2021 · 2 revisions

This tool allow to labeling images accordingly with a lot of different trainingset.

Microsoft VOTT

python tools/voc2012.py --data_dir C:\dev\yolov3-tf2\data\voc2012_raw\VOCdevkit\VOC2012 --split train --output_file C:\dev\yolov3-tf2\data\voc2012_train.tfrecord`
python tools/voc2012.py --data_dir C:\dev\yolov3-tf2\data\voc2012_raw\VOCdevkit\VOC2012 --split val --output_file C:\dev\yolov3-tf2\data\voc2012_val.tfrecord`

python train.py --dataset ./data/voc2012_train.tfrecord --val_dataset ./data/voc2012_val.tfrecord --classes ./data/voc2012.names --num_classes 20 --mode fit --transfer darknet --batch_size 16 --epochs 10 --weights ./checkpoints/yolov3.tf --weights_num_classes 80 `

  • Setup the project

    • Describe the idea
    • Identify the product shape

  • Project management

    • Define the MVP (Minimum Viable Product)
    • Identify the KPIs (Key Performance Indicators)
    • Define the process
    • Identify and define the roadmap, checkpoints, deadlines

  • Implement the solution

    • Patterns and practices
    • Technologial Stack
    • Tools, languages, libraries and services

  • Operations maintenance support and feedback

    • Strategical organization
    • Pratical implementation

Clone this wiki locally