update readme

Author: mbrostami

README.md

@@ -0,0 +1,36 @@
+# ComfyUI Textual Inversion Training
+To train textual inversion embedding directly from ComfyUI pipeline.    
+Training scripts are modified version of diffusers examples: https://github.com/huggingface/diffusers/tree/main/examples/textual_inversion
+
+## Install
+Follow the installation here:
+https://github.com/huggingface/diffusers/tree/main/examples/textual_inversion
+
+Then clone the repository inside custom_nodes:  
+```
+git clone github.com/mbrostami/ComfyUI-TITrain
+cd ComfyUI-TITrain
+pip install -r requirements.txt
+```
+
+Restart and refresh the ComfyUI.
+ 
+## Nodes
+### Textual Inversion Training SD1.5
+This node is designed to train SD 1.5 encoder (with progress bar)   
+([./screenshot.png)[img]
+
+<img alt="alt_text" src="screenshot.png" />
+
+
+## Parameters
+#### train_dir
+This is where the input images are going to be stored, if directory doesn't exist in `ComfyUI/output/` it will be created. Consider changing the value if you want to train different embeddings.    
+#### extra_prompts 
+There are list of prompts inside the training script, that is shuffled with each image to be used for training. I added this extra prompts that you can put additional prompt in each line, so randomly selected line from here will be appended to the final prompt. Usually you should put a part of prompt that you don't want to be considered as the embedding itself. (e.g. white background, close up, tattoo, etc)   
+
+
+Other parameters are explained in diffusers repository: 
+https://github.com/huggingface/diffusers/tree/main/examples/textual_inversion    
+
+

screenshot.png

@@ -390,11 +390,11 @@ def train(
         return run_script(args, max_train_steps)
 
 NODE_CLASS_MAPPINGS = {
-    "TextualInversionTrainingSDXL": TextualInversionTrainingSDXL,
+    # "TextualInversionTrainingSDXL": TextualInversionTrainingSDXL,
     "TextualInversionTraining": TextualInversionTraining
 }
 
 NODE_DISPLAY_NAME_MAPPINGS = {
-    "TextualInversionTrainingSDXL": "Textual Inversion Training SDXL",
+    # "TextualInversionTrainingSDXL": "Textual Inversion Training SDXL",
     "TextualInversionTraining": "Textual Inversion Training SD1.5"
 }