• tal@lemmy.today
    link
    fedilink
    English
    arrow-up
    11
    ·
    edit-2
    1 year ago

    First without space, second with space. Based on my single-username dataset, not sure it makes a huge difference on at least some engines. I’m guessing that if an existing term doesn’t tokenize, one way or another, substrings get used.

    ProfessorProteus

    Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 0, Size: 1024x1024, Model hash: ebf42d1fae, Model: realmixXL_v15, Token merging ratio: 0.5, Version: v1.5.1

    Professor Proteus

    Steps: 20, Sampler: Euler a, CFG scale: 7, Seed: 0, Size: 1024x1024, Model hash: ebf42d1fae, Model: realmixXL_v15, Token merging ratio: 0.5, Version: v1.5.1