A Personal Excercise for AMAP-TECH

Datasets

You can find the datasets in Tianchi.

Base Environment

• python=3.8
• cuda=12.0

How to Run

• amp_tech.ipynb is the final version.
• en_clf.ipynb is the classifier for the json.
• obj_detect.ipynb is to detect the object in the datasets with the YOLO V3.

The Tutorial

You can represent the outcom through the below tutorial.

1. The whole Pipeline
2. Feature Extraction

The Advantages

I have made the outcome of the baseline increase 10%.

Author

It's Yongce Liu, here.