Stabilize Space: switch to simple Interface mode, single text output, share=True

app.py

@@ -1,8 +1,8 @@
 import json
 import threading
 from pathlib import Path
-
-import gradio as gr
+
+import gradio as gr
 import torch
 import torch.nn as nn
 from PIL import Image
@@ -120,17 +120,11 @@ def _ensure_loaded():
 
 def predict(img):
     if img is None:
-        yield "error: no image", "status: please upload image first"
-        return
-    yield "", "status: starting"
+        return "error: no image"
     try:
-        if not _MODEL_READY:
-            yield "", "status: loading model (first run takes longer)"
         _ensure_loaded()
     except Exception:
-        yield f"error: model load failed: {_MODEL_ERR}", "status: failed"
-        return
-    yield "", "status: model ready, running inference"
+        return f"error: model load failed: {_MODEL_ERR}"
     if img.mode != "RGB":
         img = img.convert("RGB")
     proc = processor(images=img, return_tensors="pt")
@@ -140,17 +134,16 @@ def predict(img):
     pred_score = pred_01 * (score_max - score_min) + score_min
     score_int = int(round(pred_score))
     score_int = max(int(score_min), min(int(score_max), score_int))
-    yield f"score_{score_int} (raw={pred_score:.4f})", "status: done"
-
-
-with gr.Blocks() as demo:
-    gr.Markdown("# SigLIP2 Aesthetic Scorer Demo")
-    inp = gr.Image(type="pil", label="Image")
-    out = gr.Textbox(label="Result")
-    status = gr.Textbox(label="Status", value="status: idle")
-    btn = gr.Button("Predict")
-    btn.click(fn=predict, inputs=[inp], outputs=[out, status])
+    return f"score_{score_int} (raw={pred_score:.4f})"
+
 
+demo = gr.Interface(
+    fn=predict,
+    inputs=gr.Image(type="pil", label="Image"),
+    outputs=gr.Textbox(label="Result"),
+    title="SigLIP2 Aesthetic Scorer Demo",
+    description="Upload image and get score_1..score_9",
+    allow_flagging="never",
+)
 
-demo.queue(default_concurrency_limit=1)
-demo.launch(server_name="0.0.0.0", server_port=7860, show_api=False, ssr_mode=False)
+demo.launch(server_name="0.0.0.0", server_port=7860, show_api=False, ssr_mode=False, share=True)