update

config.json

@@ -40,15 +40,22 @@
   ],
   "codebook_dim": 64,
   "codebook_size": 1024,
+  "downsample_factor": 320,
   "downsample_mode": "step_down",
+  "dtype": "float32",
   "initializer_range": 0.02,
   "kernel_size": 3,
   "model_type": "higgs_audio_tokenizer",
   "num_quantizers": 8,
+  "pad": 160,
   "sample_rate": 24000,
   "semantic_model_config": {
+    "_name_or_path": "bosonai/hubert_base",
     "activation_dropout": 0.1,
     "apply_spec_augment": true,
+    "architectures": [
+      "HubertModel"
+    ],
     "attention_dropout": 0.1,
     "classifier_proj_size": 256,
     "conv_bias": false,
@@ -83,6 +90,7 @@
     "ctc_loss_reduction": "sum",
     "ctc_zero_infinity": false,
     "do_stable_layer_norm": false,
+    "dtype": "float32",
     "feat_extract_activation": "gelu",
     "feat_extract_norm": "group",
     "feat_proj_dropout": 0.0,
@@ -122,7 +130,6 @@
     2,
     4
   ],
-  "torch_dtype": "float32",
   "transformers_version": "4.56.0.dev0",
   "unit_kernel_size": 3
 }