Vishaal Udandarao

Reproducing all numbers in Tab 3

1

Hey, thanks for your great work. I am able to reproduce the CLIP-Score and ImageReward numbers for the Preference Acc. column of tab. 3 of your paper using the `test.py`...

Releasing trained topic models?

1

Hey, thanks for the great work -- do you plan to release your trained LDA model for the analysis in sec 4.2? Thanks!

Question about class thresholds

2

I am trying to run RAM++ on open-set classes. I noticed that you have two files: `ram_tag_list_threshold.txt` and `ram_tag_list.txt`. I assume these are confidence thresholds for each of the classes...

Code release?

2

Hey, thanks for the amazing, impactful work. I'm very curious to try out your method on some of my own tasks, do you have an approximate timeline of when the...

Question about LoRA alpha

1

Hi, thanks for your great work. I noticed that in your scripts, you hard-coded the lora alpha to be 128 and the rank r to be 4 (therefore leading to...

Questions about KAdaptation implementation

1

Hi, thanks for the great work and releasing the code to reproduce it. I have a few questions regarding the kronecker adaptation forward pass through the adapter modules: (1) The...

Releasing code?

1

Hey, do you have a timeline for when you will release the code and pre-trained models?

Fine-tuned checkpoint release?

2

Hi, thanks for the great work and releasing the dataset. Are you also planning to release the fine-tuned weights of the CoTBLIP model?

Release of human annotations

1

Hi, thanks for your great work and releasing your code and data publicly! I couldn't find the human annotations in this repo. In the paper you say: ``` We verify...

Dataset with human data

4

Hey, Thanks for the great work: I enjoyed reading the paper, and your proposed iterative improving strategy seems super interesting. I am interested in the human data you collected (DecomposedCaptions4k...