darkpromise

About the training script

3

Hello, a nice work for metric learning ! I'm interesting for the method and want to re-produce the results. Can you provide the training script for each dataset (e.g., CUB200...

A nice work for metric learning! I find in the backbone network, a LayerNorm is used after the final embedding/fc layer (before L2-normalization). Why use this operation? Since previous metric...

The LoRA training codes and scripts

1

A significant achievement in aligning Vision-Language Models! While running the code 'RLAIF-V/muffin/train/train_llava15.py', I noticed that all model parameters are trainable. Due to hardware limitations, could you kindly provide the LoRA...

A nice work on MLLM ! I want to know that how can I get the image mask (patch-wise mask) for every image at the inference stage. It seems that...

the query to patch attention for linear projection

1

A very impressive work for MLLM interpretability. I want to know how to compute the query-to-patch attention map (the top lines of Fig. 3) for linear projection (e.g., LLaVA), since...