Update app.py

app.py

@@ -50,24 +50,72 @@ def get_video_svd(image_in):
     print(result)
     return result[0]["video"]
 
-def infer(image_in, conditional_pose, prompt, chosen_model):
+def infer(image_in, camera_shot, conditional_pose, prompt, chosen_model):
+    if camera_shot == "custom":
+        if conditional_pose != None:
+            conditional_pose = conditional_pose
+        else :
+            raise gr.Error("No custom conditional shot found !")
+    else :
+        if camera_shot == "close-up" :
+            #
+        elif camera_shot == "medium close-up":
+            #
+        elif camera_shot == "medium shot":
+            #
+        elif camera_shot == "cowboy shot":
+            #
+        elif camera_shot == "medium full shot":
+            #
+        elif camera_shot == "full shot":
+            #
+    
     iid_img = get_instantID(image_in, conditional_pose, prompt)
+    
     if chosen_model == "i2vgen-xl" :
         video_res = get_video_i2vgen(iid_img, prompt)
     elif chosen_model == "stable-video" :
         video_res = get_video_svd(image_in)
+    
     print(video_res)
+    
     return video_res
 
-gr.Interface(
-    fn = infer,
-    inputs = [
-        gr.Image(type="filepath", label="Face to copy"),
-        gr.Image(type="filepath", label="Conditional pose (Optional)"),
-        gr.Textbox(label="Prompt"),
-        gr.Radio(label="Choose a model", choices=["i2vgen-xl", "stable-video"], value="i2vgen-xl", interactive=False, visible=False)
-    ],
-    outputs = [
-        gr.Video()
-    ]
-).launch()
+with gr.Blocks as demo:
+    with gr.Column():
+        gr.HTML("""
+        
+        """)
+        with gr.Row():
+            with gr.Column():
+                face_in = gr.Image(type="filepath", label="Face to copy")
+                camera_shot = gr.Dropdown(
+                    label = "Camera Shot", 
+                    info = "Use standard camera shots vocabulary, or drop your custom shot as conditional pose (1280*720 ratio is recommended)"
+                    choices = [
+                        "custom", "close-up", "medium close-up", "medium shot", "cowboy shot", "medium full shot", "full shot"
+                    ],
+                    value = "custom"
+                )
+                condition_shot = gr.Image(type="filepath", label="Custom conditional shot (Optional)")
+                prompt = gr.Textbox(label="Prompt")
+                chosen_model = gr.Radio(label="Choose a model", choices=["i2vgen-xl", "stable-video"], value="i2vgen-xl", interactive=False, visible=False)
+                submit_btn = gr.Button("Submit")
+            with gr.Column():
+                video_out = gr.Video()
+
+    submit_btn.click(
+        fn = infer,
+        inputs = [
+            face_in,
+            camera_shot,
+            condition_shot,
+            prompt,
+            chosen_model
+        ],
+        outputs = [
+            video_out
+        ]
+    )
+
+demo.queue(max_size=6).launch(debug=True)