Create app.py

app.py

@@ -0,0 +1,34 @@
+import gradio as gr
+import torch
+from transformers import AutoProcessor, AutoModel
+
+model_name = "MERaLiON/MERaLiON-SpeechEncoder-v1"
+
+processor = AutoProcessor.from_pretrained(model_name)
+model = AutoModel.from_pretrained(model_name)
+
+def encode_audio(audio):
+    # audio = (sample_rate, numpy array)
+    sr, data = audio
+
+    inputs = processor(
+        data,
+        sampling_rate=sr,
+        return_tensors="pt",
+        padding=True
+    )
+
+    with torch.no_grad():
+        embeddings = model(**inputs).last_hidden_state.mean(dim=1).squeeze().tolist()
+
+    return {
+        "embeddings": embeddings
+    }
+
+demo = gr.Interface(
+    fn=encode_audio,
+    inputs=gr.Audio(type="numpy", label="Upload audio"),
+    outputs=gr.JSON(label="Embeddings")
+)
+
+demo.launch()