It is not typically possible to blend models like that, since the training proce... | Hacker News

Hacker Newsnew | past | comments | ask | show | jobs | submit

		eoerl on Nov 24, 2022 \| parent \| context \| favorite \| on: Stable Diffusion 2.0 It is not typically possible to blend models like that, since the training process is (lateral) order insensitive, as far as the model goes.

liuliu on Nov 24, 2022 | [–]

I thought so too until found that there are quite a bit of literatures nowadays about "merging" weights, for example, this one: https://arxiv.org/pdf/1811.10515.pdf and also the OpenCLIP paper.

ShamelessC on Nov 24, 2022 | [–]

Is that still the case when all models have a common ancestor (i.e. finetuned) and haven’t yet overfit on new data?

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact