Update tiny models for VibeVoiceAcousticTokenizerEncoderModel

config.json

@@ -0,0 +1,26 @@
+{
+  "architectures": [
+    "VibeVoiceAcousticTokenizerEncoderModel"
+  ],
+  "channels": 1,
+  "depths": [
+    1,
+    1
+  ],
+  "downsampling_ratios": [
+    2
+  ],
+  "dtype": "float32",
+  "ffn_expansion": 4,
+  "hidden_act": "gelu",
+  "hidden_size": 32,
+  "initializer_range": 0.01,
+  "kernel_size": 3,
+  "layer_scale_init_value": 1e-06,
+  "model_type": "vibevoice_acoustic_tokenizer_encoder",
+  "n_filters": 4,
+  "num_filters": 32,
+  "rms_norm_eps": 1e-05,
+  "transformers_version": "5.16.0.dev0",
+  "vae_std": 0.625
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3eaaba609b9fce85beca378226733fb5ee3da0be77eccd93c0659c4ae484b555
+size 229536

preprocessor_config.json

@@ -0,0 +1,21 @@
+{
+  "audio_processor": {
+    "eps": 1e-06,
+    "feature_extractor_type": "VibeVoiceTokenizerProcessor",
+    "normalize_audio": true,
+    "sampling_rate": 24000,
+    "target_dB_FS": -25
+  },
+  "db_normalize": true,
+  "eps": 1e-06,
+  "feature_extractor_type": "VibeVoiceAcousticTokenizerFeatureExtractor",
+  "feature_size": 1,
+  "language_model_pretrained_name": "Qwen/Qwen2.5-1.5B",
+  "normalize_audio": true,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 24000,
+  "speech_tok_compress_ratio": 3200,
+  "target_dB_FS": -25
+}