DETR PyTorch Implementation | DETR Tutorial Part 2

In this video, we dive into the implementation of DETR (DEtection TRansformer) for object detection using PyTorch. This is Part 2 of the DETR tutorial, where using our understanding from Part 1 Video, we get to actually building and training the DETR model and see the an implementation of end-to-end object detection with transformers. We also visualize the model attention maps after training it on voc dataset. By looking through the voc dataset training code, one should get a sense of how to train DETR model on your own custom dataset and even implement your own detr from scratch.

The detr tutorial video is divided in four sections:

Recap of DETR Explanation
DETR PyTorch Implementation
Training DETR Object Detection on VOC Dataset
Discuss results and visualisation of DETR model attention heatmaps

⏱️ Timestamps
00:00 Intro
01:03 DETR Explanation Recap
07:33 Detection Transformer PyTorch Module Initialisation
17:05 Generation Prediction from DETR Layers
26:57 Matching Predictions and Target in DETR
37:52 DETR Loss Implementation
45:06 DETR Inference Code
46:57 Training DETR on VOC
52:24 Results of DETR Model Training
1:00:27 Visualisations of DETR Attention Maps
1:05:02 Outro

📖 Resources
DETR Paper - https://tinyurl.com/exai-detr-paper
DETR Explanation Video -    • DETR Explained | End-to-End Object Detecti...  
DETR Official Implementation - https://github.com/facebookresearch/detr
My DETR Implementation - https://github.com/explainingai-code/...

🔔 Subscribe :
https://tinyurl.com/exai-channel-link

Email - [email protected]