:wq

.gitignore

@@ -0,0 +1 @@
+.venv

app.py

@@ -0,0 +1,26 @@
+import gradio as gr
+import spaces
+from PIL import Image
+from transformers import PaliGemmaForConditionalGeneration, AutoProcessor
+
+TITLE = "E621 Tagger"
+DESCRIPTION = "Tag images with E621 tags"
+
+MODEL_ID = "estrogen/paligemma2-3b-e621-224"
+
+model = PaliGemmaForConditionalGeneration.from_pretrained(MODEL_ID)
+processor = AutoProcessor.from_pretrained(MODEL_ID)
+
+@spaces.GPU
+def tag_image(image):
+    inputs = processor(images=image, text="<image>tag en", return_tensors="pt")
+    outputs = model.generate(**inputs)
+    return processor.decode(outputs[0], skip_special_tokens=True)
+
+gr.Interface(
+    fn=tag_image,
+    inputs=gr.Image(type="pil"),
+    outputs=gr.Textbox(type="text"),
+    title=TITLE,
+    description=DESCRIPTION,
+).launch()