Update app.py

app.py

@@ -37,7 +37,14 @@ def save_media_and_get_path(media, media_type):
 
 
 @spaces.GPU
-def qwen_inference(media, media_type, text_input=None):
+def qwen_inference(media, text_input=None):
+    if isinstance(media, Image.Image):
+        media_type = "image"
+    elif isinstance(media, str) and media.endswith((".mp4", ".webm", ".avi")):  # Check if it's a video path
+        media_type = "video"
+    else:
+        raise ValueError("Unsupported media type. Please upload an image or video.")
+
     media_path = save_media_and_get_path(media, media_type)
     
     messages = [
@@ -81,28 +88,15 @@ css = """
 with gr.Blocks(css=css) as demo:
     gr.Markdown(DESCRIPTION)
 
-    with gr.Tab(label="Image Input"):
-        with gr.Row():
-            with gr.Column():
-                input_img = gr.Image(label="Input Picture", type="pil")
-                text_input_image = gr.Textbox(label="Question")
-                submit_btn_image = gr.Button(value="Submit")
-            with gr.Column():
-                output_text_image = gr.Textbox(label="Output Text")
-        type = gr.State("image")
-
-        submit_btn_image.click(qwen_inference, [input_img, type, text_input_image], [output_text_image])
-
-    with gr.Tab(label="Video Input"):
+    with gr.Tab(label="Image/Video Input"):
         with gr.Row():
             with gr.Column():
-                input_video = gr.Video(label="Input Video")
-                text_input_video = gr.Textbox(label="Question")
-                submit_btn_video = gr.Button(value="Submit")
+                input_media = gr.Upload(label="Upload Image or Video", type="file") 
+                text_input = gr.Textbox(label="Question")
+                submit_btn = gr.Button(value="Submit")
             with gr.Column():
-                output_text_video = gr.Textbox(label="Output Text")
-        type = gr.State("video")
+                output_text = gr.Textbox(label="Output Text")
 
-        submit_btn_video.click(qwen_inference, [input_video, type, text_input_video], [output_text_video])
+        submit_btn.click(qwen_inference, [input_media, text_input], [output_text])
 
 demo.launch(debug=True)