Initial

README.md

@@ -1,3 +1 @@
----
-license: mit
----
+Bark inference endpoint

handler.py

@@ -0,0 +1,21 @@
+from bark import generate_audio, preload_models
+import torch
+import numpy as np
+import scipy
+import io
+
+preload_models()
+
+def infer(inputs):
+    text = inputs.get("text", "")
+    if not text.strip():
+        return {"error": "No input text provided."}
+
+    audio_array = generate_audio(text, history_prompt="v2/ru_speaker_0")
+
+    # Сохраняем в wav-поток
+    byte_io = io.BytesIO()
+    scipy.io.wavfile.write(byte_io, rate=24000, data=audio_array)
+    byte_io.seek(0)
+
+    return {"audio": byte_io.read()}

requirements.txt

@@ -0,0 +1,9 @@
+torch==2.5.1
+torchaudio==2.5.1
+torchvision==0.20.1
+numpy
+scipy
+pydub
+srt
+certifi
+git+https://github.com/suno-ai/bark.git