Create app.py

app.py

@@ -0,0 +1,25 @@
+import gradio as gr
+import torch
+import librosa
+from transformers import AutoProcessor, ClapModel
+
+# Load model globally for speed
+model_id = "laion/clap-htsat-unfused"
+processor = AutoProcessor.from_pretrained(model_id)
+model = ClapModel.from_pretrained(model_id)
+
+def get_vibe_vector(audio_path):
+    # Load and resample to 48kHz (required for CLAP)
+    y, sr = librosa.load(audio_path, sr=48000, duration=30)
+    inputs = processor(audios=y, return_tensors="pt", sampling_rate=48000)
+    with torch.no_grad():
+        embeddings = model.get_audio_features(**inputs)
+    return embeddings.numpy().flatten().tolist()
+
+# Define the Gradio Interface
+demo = gr.Interface(
+    fn=get_vibe_vector, 
+    inputs=gr.Audio(type="filepath"), 
+    outputs="json"
+)
+demo.launch()