ys-zong / VL-ICL

Code for paper: VL-ICL Bench: The Devil in the Details of Benchmarking Multimodal In-Context Learning
https://ys-zong.github.io/VL-ICL/
28 stars 2 forks source link

how to use dataset #2

Closed whitesockcat closed 6 months ago

whitesockcat commented 6 months ago

Good job, your work inspire me a lot.

Here is some little question about dataset.

  1. I don't see the train test split from dataset link.
  2. I don't know how to orgnize in contex sample from dataset. There seems is only sample one by one, not few shot.

Looking forward to get your responce!

whitesockcat commented 6 months ago

Good job, your work inspire me a lot.

Here is some little question about dataset.

  1. I don't see the train test split from dataset link.
  2. I don't know how to orgnize in contex sample from dataset. There seems is only sample one by one, not few shot.

Looking forward to get your responce!

I got it, query.json is test split. support.json is train split

whitesockcat commented 6 months ago

Good job, your work inspire me a lot.

Here is some little question about dataset.

  1. I don't see the train test split from dataset link.
  2. I don't know how to orgnize in contex sample from dataset. There seems is only sample one by one, not few shot.

Looking forward to get your responce!

https://github.com/ys-zong/VL-ICL/blob/main/utils/ICL_utils.py here is data using solution.

Good job!

ys-zong commented 6 months ago

Yes, exactly.