video datasets

Former-commit-id: 8cafc7b055

src/llamafactory/webui/components/chatbot.py

@@ -43,8 +43,12 @@ def create_chat_box(
                         system = gr.Textbox(show_label=False)
                         tools = gr.Textbox(show_label=False, lines=3)
 
-                    with gr.Column() as image_box:
-                        image = gr.Image(sources=["upload"], type="pil")
+                    with gr.Column() as mm_box:
+                        with gr.Tab("Image"):
+                            image = gr.Image(sources=["upload"], type="pil")
+
+                        with gr.Tab("Video"):
+                            video = gr.Video(sources=["upload"])
 
                 query = gr.Textbox(show_label=False, lines=8)
                 submit_btn = gr.Button(variant="primary")
@@ -63,7 +67,7 @@ def create_chat_box(
         [chatbot, messages, query],
     ).then(
         engine.chatter.stream,
-        [chatbot, messages, system, tools, image, max_new_tokens, top_p, temperature],
+        [chatbot, messages, system, tools, image, video, max_new_tokens, top_p, temperature],
         [chatbot, messages],
     )
     clear_btn.click(lambda: ([], []), outputs=[chatbot, messages])
@@ -76,8 +80,9 @@ def create_chat_box(
             role=role,
             system=system,
             tools=tools,
-            image_box=image_box,
+            mm_box=mm_box,
             image=image,
+            video=video,
             query=query,
             submit_btn=submit_btn,
             max_new_tokens=max_new_tokens,

src/llamafactory/webui/components/infer.py

@@ -68,7 +68,7 @@ def create_infer_tab(engine: "Engine") -> Dict[str, "Component"]:
     engine.manager.get_elem_by_id("top.model_name").change(
         lambda model_name: gr.Column(visible=get_visual(model_name)),
         [engine.manager.get_elem_by_id("top.model_name")],
-        [chat_elems["image_box"]],
+        [chat_elems["mm_box"]],
     )
 
     return elem_dict