From 44d4b0714c662e5780fb0fed6be682d5851fee92 Mon Sep 17 00:00:00 2001
From: AUTOMATIC1111 <16777216c@gmail.com>
Date: Tue, 18 Oct 2022 08:15:28 +0300
Subject: [PATCH] Updated Textual Inversion (markdown)

---
 Textual-Inversion.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Textual-Inversion.md b/Textual-Inversion.md
index cda6ec6..3a94688 100644
--- a/Textual-Inversion.md
+++ b/Textual-Inversion.md
@@ -61,7 +61,7 @@ This takes images from a directory, processes them to be ready for textual inver
 
 ### Training an embedding
 - **Embedding**: select the embedding you want to train from this dropdown.
-- **Learning rate**: how fast should the training go. The danger of setting this parameter to a high value is that you may break the embedding if you set it too high. If you see `Loss: nan` in the training info textbox, that means you failed and the embedding is dead. With the default value, this should not happen.
+- **Learning rate**: how fast should the training go. The danger of setting this parameter to a high value is that you may break the embedding if you set it too high. If you see `Loss: nan` in the training info textbox, that means you failed and the embedding is dead. With the default value, this should not happen. It's possible to specify multiple learning rates in this settins using the following syntax: `0.005:100, 1e-3:1000, 1e-5` - this will train with lr of `0.005` for first 100 steps, then `1e-3` until 1000 steps, then `1e-5` until the end.
 - **Dataset directory**: directory with images for training. They all must be square.
 - **Log directory**: sample images and copies of partially trained embeddings will be written to this directory.
 - **Prompt template file**: text file with prompts, one per line, for training the model on. See files in directory `textual_inversion_templates` for what you can do with those. Use `style.txt` when training styles, and `subject.txt` when training object embeddings. Following tags can be used in the file: