lllyasviel / lllyasviel/ControlNet

Question about construction of training datasets

Open
#317 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
34.1k
Forks
3k
PR merge metrics
No merged PRs in 30d

Description

Hello, I have read the part of your paper regarding the synthesis of training datasets. I noticed that you mentioned obtaining 164K segmentation-image-caption pairs from the ADE20K dataset using image caption methods. However, I was a bit confused as ADE20K only contains approximately 25,574 images. I was wondering if you could kindly provide me with further details about the construction of the training dataset with 164K pairs? I appreciate your time and thank you in advance for your assistance.

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with the paper section describing synthesis of the training datasets and the ADE20K reference. Trace how the stated 164K segmentation-image-caption pairs were obtained from approximately 25,574 images. Done means documenting a clear, supported explanation of the dataset construction and resolving the discrepancy raised in the issue.

Written by the indexing model from the issue text.

Assessment

Domain
data, documentation, machine-learning
Issue type
Documentation
Difficulty
4/5
Estimated time
3-5 days
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.