CodeCompass
CodeCompass copied to clipboard
A Repository chatbot and Recommendation system for Github users.
A great enhancement for the "CodeCompass" project would be to introduce user-specific model customization. By analyzing a user's past GitHub activity (e.g., commits, issues, repositories followed), the system could fine-tune...
@miguelgfierro Please check your IE email. We will start preparing our presentation soon, test run on Friday the 9th and official presentation the following friday.
Redis
The model takes 20-30s to generate the recommendations. It shouldn't take that long, try to optimize.
Implement other models apart from lightgbm. You can try faster models like logistic regression, review VW, or learn. In Recommenders you have those and many more examples.
Implement other chatbots, in particular, llama3 just released a few days ago, it would be nice.
One way to create a solution that is closer to a real product is to create a batch architecture. You can create a pipeline that gets the new github data,...