More Thoughts on Mamba: Predicting the view count of a Youtube video from its thumbnail with VisionMamba

Follow the full discussion on Reddit.
For fun, I wrote an article about my project to predict the view count of a video from its thumbnail, using a variation of Vision Transformer that uses Mamba as a backbone instead. It's a really nice model to use. Its memory-lightness means I can use it as a GPU-poor person, which is awesome. This is the first post in a series, I plan to scale up the dataset in the future, incorporate text embeddings, and then finally use it with RLHF to finetune a stable diffusion model.