Upload folder using huggingface_hub

config.json

@@ -1,8 +1,8 @@
 {
   "architectures": [
-    "SileroVAD"
+    "AutoModelForAudioClassification"
   ],
-  "model_type": "silero_vad",
+  "model_type": "audio-classification",
   "id2label": {
     "0": "no_speech",
     "1": "speech"
@@ -11,11 +11,8 @@
     "no_speech": 0,
     "speech": 1
   },
+  "hidden_size": 768,
+  "num_labels": 2,
   "sampling_rate": 16000,
-  "threshold": 0.5,
-  "min_speech_duration_ms": 250,
-  "min_silence_duration_ms": 100,
-  "window_size_samples": 512,
-  "speech_pad_ms": 30,
   "transformers_version": "4.30.2"
 }

onnx_config.json

@@ -0,0 +1,36 @@
+{
+  "input_names": ["input"],
+  "output_names": ["logits"],
+  "dynamic_axes": {
+    "input": {
+      "0": "batch_size",
+      "1": "sequence_length"
+    },
+    "logits": {
+      "0": "batch_size",
+      "1": "num_classes"
+    }
+  },
+  "opset_version": 14,
+  "supported_backends": ["onnxruntime", "webgl", "wasm"],
+  "quantized": false,
+  "model_inputs": [
+    {
+      "name": "input",
+      "shape": [1, -1],
+      "dtype": "float32"
+    }
+  ],
+  "model_outputs": [
+    {
+      "name": "logits",
+      "shape": [1, 2],
+      "dtype": "float32"
+    }
+  ],
+  "pipeline_tag": "audio-classification",
+  "metadata": {
+    "transformers_version": "4.30.2",
+    "framework": "onnx"
+  }
+}

preprocessor_config.json

@@ -1,5 +1,5 @@
 {
-  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_extractor_type": "AudioFeatureExtractor",
   "padding_side": "right",
   "sampling_rate": 16000,
   "return_attention_mask": true,

tokenizer_config.json

@@ -0,0 +1,11 @@
+{
+  "name_or_path": "facebook/wav2vec2-base",
+  "return_attention_mask": true,
+  "do_normalize": true,
+  "padding": "max_length",
+  "max_length": 16000,
+  "sampling_rate": 16000,
+  "feature_size": 1,
+  "padding_value": 0.0,
+  "truncation": true
+}