update app

app.py

@@ -12,15 +12,16 @@ with gr.Blocks() as demo:
     ### Input A FACE and get the body index
     """
     )
-    image = gr.Image(type="pil")
-    with gr.Column(scale=1, min_width=600):
-        # 設定輸出元件
-        heights = gr.Textbox(label="Heignt")
-        bust = gr.Textbox(label="Bust")
-        waist = gr.Textbox(label="Waist")
-        hips = gr.Textbox(label="Hips")
-        en_des = gr.Textbox(label="English description")
-        zh_des = gr.Textbox(label="Chinese description")
+    with gr.Row():
+        image = gr.Image(type="pil")
+        with gr.Column(scale=1, min_width=600):
+            # 設定輸出元件
+            heights = gr.Textbox(label="Heignt")
+            bust = gr.Textbox(label="Bust")
+            waist = gr.Textbox(label="Waist")
+            hips = gr.Textbox(label="Hips")
+            en_des = gr.Textbox(label="English description")
+            zh_des = gr.Textbox(label="Chinese description")
     
     #設定按鈕
     submit = gr.Button("Submit")

predictor.py

@@ -224,7 +224,7 @@ def inference(inp_img, classes = ['big', 'small'], epoch = 6):
         idx = preds.numpy()[0]
 
         # unconditional image captioning
-        inputs = processor(inp_img, return_tensors="pt")
+        inputs = processor(inp_img, return_tensors="pt").to(device)
         out = model_blip.generate(**inputs)
         description = processor.decode(out[0], skip_special_tokens=True)
         description_tw = translator.translate(description)