Hi, I am looking to replicate the work done as part of this repo, can you please share the training data and model weights?