Create config.json

config.json

@@ -0,0 +1,41 @@
+{
+    "model": "xtts_v2",
+    "output_path": "./output_model",
+    "audio": {
+        "data_path": "./voxpopuli_es_500/wav_data",
+        "meta_file_train": "./voxpopuli_es_500/metadata.json",
+        "sample_rate": 22050,
+        "num_mels": 80,
+        "max_wav_value": 32768
+    },
+    "training": {
+        "epochs": 5,
+        "batch_size": 4,
+        "save_step": 100,
+        "eval_step": 100,
+        "lr": 0.0001,
+        "optimizer": "adam",
+        "scheduler": "NoamLR",
+        "use_cuda": true,
+        "mixed_precision": true,
+        "seed": 1234
+    },
+    "datasets": [
+        {
+            "formatter": "coqui",
+            "meta_file_train": "./voxpopuli_es_500/metadata.json",
+            "path": "./voxpopuli_es_500/wav_data"
+        }
+    ],
+    "model_params": {
+        "decoder_input_dim": 1024,
+        "decoder_hidden_dim": 1024,
+        "dropout": 0.1,
+        "num_heads": 8,
+        "num_layers": 6,
+        "speaker_embedding_dim": 256,
+        "style_embedding_dim": 256,
+        "language_embedding_dim": 64,
+        "max_text_len": 400
+    }
+}