thomasly / contextsub Goto Github PK
View Code? Open in Web Editor NEWUse ContextPred method to pretrain GNN models on subgraph embeddings
License: MIT License
Use ContextPred method to pretrain GNN models on subgraph embeddings
License: MIT License
Use lightGBM with molecular descriptors to predict BBB penetrate capability of drugs. It can be used for scaffold splitting.
Add the external testing set with 74 external molecules into the loader.py
precompute batch indicator for each molecule and its substructures and save them to Data object so pyg can use it to compute batch pooling indicator fast.
compare the current version of comtextSub (no structure channel) to contextPred
Evaluate the pretraining result by analyzing the embeddings of the same substructure with different contexts.
Compute atom partial charges and add them as atom properties for model pretraining and finetuning.
Fix the output layer of the GNN model to use substructure embeddings as outputs.
Add the function option to finetune the pretrained model based on substructure embeddings.
Update the dataloader.py to include the new data set support.
enable weights freezing during fine-tuning.
Freeze the ContextPred and ContextSub weights separately. Specifically, unfreeze the weights of contextSub first during fine-tuning and then unfreeze the weights of ContextPred. See if this strategy can improve the final performance.
Add unit test for new added functions.
Convert the source code from ContextPred paper (https://github.com/snap-stanford/pretrain-gnns) to learn subgraph embeddings.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.