Skip to content

Train SALAD on own dataser #10

@Skwarson96

Description

@Skwarson96

I wanted to repeat the steps in the README.md to train the model on my own dataset. First, I ran the create_fg_masks.py script, and everything was fine. Then I tried the create_pseudo_labels.py script, and I got the error:

ValueError: Found array with 0 sample(s) (shape=(0, 768)) while a minimum of 1 is required by KMeans

I assume the problem is with the paths. I suspect the problem might be here:

root_dir=data_path, bg_dir=mask_path, category="/train/good", resize_shape=512

Could you tell me what the directory structures should look like at each stage?

Thank you

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions