IntelLabs / IntelLabs/FloorSet

Training Set have MIB group which contain different area target

Open
#12 2 comments 1 reaction 0 assignees View on GitHub
Dominant language
Jupyter Notebook
Stars
119
Forks
35
PR merge metrics
No merged PRs in 30d

Description

Dear IntelLabs and ICCAD 2026 FloorSet organizers,

I have found that in the training set in https://huggingface.co/datasets/IntelLabs/FloorSet have tons of entry have invalid MIB groups.

By using following script to detect/reproduce the problem,
https://gist.github.com/KohakuBlueleaf/184911ab63d5a4a0a48b5a6cf7134e3a

We can find there is lot of row that have "same MIB groups id" while only the first row contain the "real MIB groups". All other row result in "the blocks that is noting in MIG group actually have different target area" which means "achiving MIB with same H,W is impossible"

We want to know if this is intended? if so, does MIB still a scoring target?

If this is not intended, will there be any updated/corrected training set?

Contributor guide

No contributing guide indexed for this repository

Research direction

Start with the Hugging Face FloorSet training set and run the detection script linked in the issue. Inspect rows sharing an MIB group ID and compare their real MIB groups and target areas. Done means determining whether these differing areas are intended and documenting or correcting the affected dataset entries accordingly.

Written by the indexing model from the issue text.

Assessment

Tech stack
huggingface
Domain
data, machine-learning
Issue type
Bug
Difficulty
3/5
Estimated time
1-2 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.