Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -52,7 +52,7 @@ BADGE = """
|
|
52 |
"""
|
53 |
|
54 |
LOGO = '<p align="center"><img width="350" src="https://raw.githubusercontent.com/yeliudev/VideoMind/refs/heads/main/.github/logo.png"></p>'
|
55 |
-
DISC = '**VideoMind** is a multi-modal agent framework that enhances video reasoning by emulating *human-like* processes, such as *breaking down tasks*, *localizing and verifying moments*, and *synthesizing answers*.
|
56 |
|
57 |
# yapf:disable
|
58 |
EXAMPLES = [
|
@@ -624,7 +624,7 @@ def build_demo():
|
|
624 |
submit_ctx = submit_ctx.then(main, [video, prompt, role, temperature, max_new_tokens], chat)
|
625 |
submit_ctx.then(enable_btns, None, [random_btn, reset_btn, submit_btn])
|
626 |
|
627 |
-
gr.Markdown('
|
628 |
|
629 |
with gr.Column(scale=5):
|
630 |
chat.render()
|
|
|
52 |
"""
|
53 |
|
54 |
LOGO = '<p align="center"><img width="350" src="https://raw.githubusercontent.com/yeliudev/VideoMind/refs/heads/main/.github/logo.png"></p>'
|
55 |
+
DISC = '**VideoMind** is a multi-modal agent framework that enhances video reasoning by emulating *human-like* processes, such as *breaking down tasks*, *localizing and verifying moments*, and *synthesizing answers*. Our method addresses the unique challenges of temporal-grounded reasoning in a progressive strategy. This demo showcases how VideoMind-2B handles video-language tasks. Please open an <a href="https://github.com/yeliudev/VideoMind/issues/new" target="_blank">issue</a> if you meet any problems.' # noqa
|
56 |
|
57 |
# yapf:disable
|
58 |
EXAMPLES = [
|
|
|
624 |
submit_ctx = submit_ctx.then(main, [video, prompt, role, temperature, max_new_tokens], chat)
|
625 |
submit_ctx.then(enable_btns, None, [random_btn, reset_btn, submit_btn])
|
626 |
|
627 |
+
gr.Markdown('###### Need example data? Explore examples tab or click 🔮 Random to sample one!')
|
628 |
|
629 |
with gr.Column(scale=5):
|
630 |
chat.render()
|