Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -62,7 +62,7 @@ for filename, url in sample_videos_urls.items():
|
|
62 |
print(f"{filename} already exists. Skipping download.")
|
63 |
|
64 |
csv.field_size_limit(100000000)
|
65 |
-
options = ['language', "
|
66 |
load_size = 224
|
67 |
plot_size = 224
|
68 |
|
@@ -74,16 +74,12 @@ video_output2 = gr.Video(label="Multi-Head Audio Video Attention (Only Availible
|
|
74 |
height=480)
|
75 |
video_output3 = gr.Video(label="Visual Features", height=480)
|
76 |
|
77 |
-
|
78 |
-
models = {o: None for o in options}
|
79 |
-
|
80 |
-
for o in options:
|
81 |
-
print(o)
|
82 |
-
models[o] = LitAVAligner.from_pretrained(f"mhamilton723/DenseAV-{o}")
|
83 |
|
84 |
def process_video(video, model_option):
|
85 |
# model = models[model_option].cuda()
|
86 |
model = models[model_option]
|
|
|
87 |
|
88 |
original_frames, audio, info = torchvision.io.read_video(video, end_pts=10, pts_unit='sec')
|
89 |
sample_rate = 16000
|
|
|
62 |
print(f"{filename} already exists. Skipping download.")
|
63 |
|
64 |
csv.field_size_limit(100000000)
|
65 |
+
options = ['language', "sound_and_language", "sound"] # Changed second to initial
|
66 |
load_size = 224
|
67 |
plot_size = 224
|
68 |
|
|
|
74 |
height=480)
|
75 |
video_output3 = gr.Video(label="Visual Features", height=480)
|
76 |
|
77 |
+
models = {o: LitAVAligner.from_pretrained(f"mhamilton723/DenseAV-{o}") for o in ['language', "sound-language", "sound"]}
|
|
|
|
|
|
|
|
|
|
|
78 |
|
79 |
def process_video(video, model_option):
|
80 |
# model = models[model_option].cuda()
|
81 |
model = models[model_option]
|
82 |
+
print(model_option)
|
83 |
|
84 |
original_frames, audio, info = torchvision.io.read_video(video, end_pts=10, pts_unit='sec')
|
85 |
sample_rate = 16000
|