From 89c823a56ca3a1e28589542e789b87286a5e5cf2 Mon Sep 17 00:00:00 2001 From: AUTOMATIC1111 <16777216c@gmail.com> Date: Sat, 31 Dec 2022 21:51:03 +0300 Subject: [PATCH] Updated Features (markdown) --- Features.md | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-) diff --git a/Features.md b/Features.md index d7185bd..c75c37f 100644 --- a/Features.md +++ b/Features.md @@ -4,12 +4,14 @@ All examples are non-cherrypicked unless specified otherwise. # Alt-Diffusion -- download the checkpoint (from here: https://drive.filen.io/d/be043c9e-a171-4356-a749-e7840dfcb67e#Z21U5JO7HE5g1fSa7kY5nSZI2dInaUHw) -- put it into models/Stable-Diffusion directory +Model trained to accept inputs in different languages. More info: https://arxiv.org/abs/2211.06679 + +- download the checkpoint from [drive.filen.io](https://drive.filen.io/d/be043c9e-a171-4356-a749-e7840dfcb67e#Z21U5JO7HE5g1fSa7kY5nSZI2dInaUHw) +- put it into `models/Stable-Diffusion` directory - grab the config from [`configs/alt-diffusion-inference.yaml`](https://github.com/AUTOMATIC1111/stable-diffusion-webui/blob/master/configs/alt-diffusion-inference.yaml) and put it into same place as the checkpoint, renaming it to have same filename (i.e. if your checkpoint is named `ad.ckpt`, the config should be named `ad.yaml`) - select the new checkpoint from the UI -Mechanically, Attention/emphasis mechanism (see below in features) is supported, but seems to have much less effect, probably due to how Alt-Diffusion is implemented. Clip skip is not supported, the setting is ignored. +Mechanically, attention/emphasis mechanism (see below in features) is supported, but seems to have much less effect, probably due to how Alt-Diffusion is implemented. Clip skip is not supported, the setting is ignored. See the PR for more info: https://github.com/AUTOMATIC1111/stable-diffusion-webui/pull/5238 @@ -18,7 +20,7 @@ See the PR for more info: https://github.com/AUTOMATIC1111/stable-diffusion-webu Models are supported: 768-v-ema.ckpt ([model](https://huggingface.co/stabilityai/stable-diffusion-2/blob/main/768-v-ema.ckpt), [config](https://raw.githubusercontent.com/Stability-AI/stablediffusion/main/configs/stable-diffusion/v2-inference-v.yaml)) and 512-base-ema.ckpt ([model](https://huggingface.co/stabilityai/stable-diffusion-2-base/blob/main/512-base-ema.ckpt), [config](https://raw.githubusercontent.com/Stability-AI/stablediffusion/main/configs/stable-diffusion/v2-inference.yaml)). 2.1 checkpoints should also work. - download the checkpoint (from here: https://huggingface.co/stabilityai/stable-diffusion-2) -- put it into models/Stable-Diffusion directory +- put it into `models/Stable-Diffusion` directory - grab the config from SD2.0 repository and put it into same place as the checkpoint, renaming it to have same filename (i.e. if your checkpoint is named `768-v-ema.ckpt`, the config should be named `768-v-ema.yaml`) - select the new checkpoint from the UI