Spaces:
Running
on
Zero
Running
on
Zero
Add title and instruction
Browse files
app.py
CHANGED
@@ -149,6 +149,23 @@ if __name__ == "__main__":
|
|
149 |
)
|
150 |
|
151 |
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
152 |
with gr.Row():
|
153 |
with gr.Column(scale=4):
|
154 |
with gr.Row():
|
|
|
149 |
)
|
150 |
|
151 |
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
152 |
+
gr.Markdown("""# DanTagGen beta DEMO
|
153 |
+
#### What is this:
|
154 |
+
DanTagGen(Danbooru Tag Generator) is a LLM model designed for generating Danboou Tags with provided informations.
|
155 |
+
It aims to provide user a more convinient way to make prompts for Text2Image model which is trained on Danbooru datasets.
|
156 |
+
|
157 |
+
#### How to use it:
|
158 |
+
1. Fill the informations on the left most section.
|
159 |
+
2. Add the general tags you want to use. ("prompt before refined")
|
160 |
+
3. If you want to ban some tags. Put them into the "black list" text area.
|
161 |
+
4. Choose the target length: **Long or Short is recommended**
|
162 |
+
* Very Short: around 10 tags
|
163 |
+
* Short: around 20 tags
|
164 |
+
* Long: around 40 tags
|
165 |
+
* very long: around 60 tags
|
166 |
+
5. Submit!!
|
167 |
+
6. You will get formated result on the upper-right section, LLM raw result on the bottom-right section.
|
168 |
+
""")
|
169 |
with gr.Row():
|
170 |
with gr.Column(scale=4):
|
171 |
with gr.Row():
|