I looked into the implementation of graph attention networks in DGL. DGL is a library for deep learning on graphs. I also talked about how data is processed inside graph attention layers.
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
🔗 notes+code: https://mashaan14.github.io/YouTube-channel/graph_neural_networks/2024_05_13_DGL_GAT
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
- website https://mashaan14.github.io/mashaan/
- github https://github.com/mashaan14
- X https://twitter.com/mashaan_14
- linkedin https://linkedin.com/in/mashaan
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
- sponsor this channel on GitHub Sponsors:
https://github.com/sponsors/mashaan14
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
📹 Video edit: Adobe Premiere Rush
🎧 Audio enhancement: Adobe Podcast
🖼️ Thumbnails: GIMP
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
Chapters:
00:00 start
00:11 DGL or PyG?
01:23 acknowledgment
01:46 GATConv layer in DGL
03:11 GATConv source code
05:00 start of the code
05:13 GAT layers from DGL
06:20 how data is processed inside GAT
07:35 breaking down the 1st layer
08:48 why concatenation not averaging?
09:41 2nd GAT layer
10:23 evaluate and train functions
10:58 importing Cora dataset
11:31 initialize the model
11:49 training and testing
▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
#graphneuralnetwork #DGL #GNN #GAT #attention #graphconvolution #pytorch-geometric #jraph #graph #GCN #pytorchgeometric #GCNConv #GATConv #DeepLearningTutorial #MachineLearningProject #AIResearch #CodingTutorial