Upload vibevoice_pipeline_config.json with huggingface_hub

vibevoice_pipeline_config.json

@@ -0,0 +1,52 @@
+{
+  "name": "VibeVoice-1.5B-CoreML",
+  "version": "1.0.0",
+  "description": "VibeVoice TTS model converted to CoreML format",
+  "components": {
+    "acoustic_encoder": {
+      "path": "vibevoice_acoustic_encoder.mlpackage",
+      "available": true
+    },
+    "semantic_encoder": {
+      "path": "vibevoice_semantic_encoder.mlpackage",
+      "available": true
+    },
+    "acoustic_connector": {
+      "path": "vibevoice_acoustic_connector.mlpackage",
+      "available": true
+    },
+    "semantic_connector": {
+      "path": "vibevoice_semantic_connector.mlpackage",
+      "available": true
+    },
+    "diffusion_head": {
+      "path": "vibevoice_diffusion_head.mlpackage",
+      "available": true
+    }
+  },
+  "inference": {
+    "audio": {
+      "sample_rate": 24000,
+      "channels": 1,
+      "downsample_factor": 3200,
+      "frame_rate_hz": 7.5,
+      "fixed_encoder_samples": 24000,
+      "fixed_encoder_note": "Acoustic and Semantic encoders accept only 24000 samples (1 sec at 24kHz); trim or pad input."
+    },
+    "diffusion": {
+      "num_steps": 20,
+      "beta_schedule": "cosine",
+      "prediction_type": "v_prediction",
+      "guidance_scale": 1.0
+    },
+    "generation": {
+      "max_audio_length_seconds": 60,
+      "max_seq_length": 4096
+    }
+  },
+  "platform": {
+    "minimum_ios_version": "17.0",
+    "minimum_macos_version": "14.0",
+    "compute_units": "ALL"
+  }
+}