Update

app.py

@@ -38,19 +38,12 @@ def sample_frame_indices(
 
 # @st.cache_resource
 @st.experimental_singleton
-def load_model():
-    feature_extractor = AutoFeatureExtractor.from_pretrained(
-        "MCG-NJU/videomae-base-finetuned-kinetics"
-    )
-    model = TimesformerForVideoClassification.from_pretrained(
-        "facebook/timesformer-base-finetuned-k400"
-    )
+def load_model(model_name: str):
+    feature_extractor = AutoFeatureExtractor.from_pretrained(model_name)
+    model = TimesformerForVideoClassification.from_pretrained(model_name)
     return feature_extractor, model
 
 
-feature_extractor, model = load_model()
-
-
 def inference(file_path: str):
     videoreader = VideoReader(VIDEO_TMP_PATH, num_threads=1, ctx=cpu(0))
 
@@ -96,6 +89,21 @@ with st.expander("INTRODUCTION"):
     """
     )
 
+model_name = st.selectbox(
+    "model_name",
+    (
+        "facebook/timesformer-base-finetuned-k400",
+        "facebook/timesformer-base-finetuned-k600",
+        "facebook/timesformer-base-finetuned-ssv2",
+        "facebook/timesformer-hr-finetuned-k600",
+        "facebook/timesformer-hr-finetuned-k400",
+        "facebook/timesformer-hr-finetuned-ssv2",
+        "fcakyon/timesformer-large-finetuned-k400",
+        "fcakyon/timesformer-large-finetuned-k600",
+    ),
+)
+feature_extractor, model = load_model(model_name)
+
 VIDEO_TMP_PATH = os.path.join("tmp", "tmp.mp4")
 uploadedfile = st.file_uploader("Upload file", type=["mp4"])