From 3fcff4ca1a8ee63a6a835be4d19871f78cd15bd0 Mon Sep 17 00:00:00 2001
From: ClashSAN <98228077+ClashSAN@users.noreply.github.com>
Date: Thu, 24 Aug 2023 09:46:33 -0400
Subject: [PATCH] Add sdxl optimizations (+feel free to add more)

---
 Optimum-SDXL-Usage.md | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)
 create mode 100644 Optimum-SDXL-Usage.md

diff --git a/Optimum-SDXL-Usage.md b/Optimum-SDXL-Usage.md
new file mode 100644
index 0000000..f0ce09e
--- /dev/null
+++ b/Optimum-SDXL-Usage.md
@@ -0,0 +1,20 @@
+Here's a quick listing of things to tune for your setup:
+
+## Commandline arguments:
+
+- (nvidia) 8gb `--medvram-sdxl --xformers`
+- (nvidia) 12gb+ `--xformers`
+- (nvidia) 4gb `--lowvram --xformers`
+
+
+## System:
+- downgrade nvidia drivers to 531 or lower prevent extreme slowdowns for largest pictures
+- add a pagefile to prevent failure to load weights due to low cpu ram
+- (Linux) install tcmalloc - greatly reducing ram usage: `sudo apt install --no-install-recommends google-perftools` [#10117](https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/10117)
+- use an SSD, for faster loadtime, especially if a pagefile is required
+- converting `.safetensors` to `.ckpt` for reduced ram usage [#12086](https://github.com/AUTOMATIC1111/stable-diffusion-webui/issues/12086#issuecomment-1691154698)
+
+## Model weights:
+
+- use vae that will not need to run in fp32  for increased speed and less vram usage: [sdxl_vae.safetensors](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/blob/main/sdxl_vae.safetensors)
+- use fp16 (~7gb) weights for less cpu ram usage
\ No newline at end of file