Update config.json

config.json

@@ -1,44 +1,41 @@
 {
-  "model": "xtts_v2",
+  "model": "xtts",
+  "run_name": "xttsv2_finetune_es",
   "output_path": "./output_model",
-  "restore_path": null,
   "datasets": [
     {
-      "formatter": "coqui",
+      "formatter": "json",
       "meta_file_train": "./voxpopuli_es_500/metadata.json",
       "path": "./voxpopuli_es_500/wav_data",
       "language": "es"
     }
   ],
-  "trainer": {
-    "max_epochs": 5,
-    "batch_size": 4,
-    "save_step": 100,
-    "log_step": 10,
-    "eval_step": 100,
-    "mixed_precision": true,
-    "num_loader_workers": 4,
-    "eval_split_max_size": 50,
-    "run_eval": true
-  },
-  "audio": {
-    "sample_rate": 22050,
-    "win_length": 1024,
-    "hop_length": 256,
-    "fft_size": 1024,
-    "mel_fmin": 0,
-    "mel_fmax": 8000
-  },
-  "optimizer": {
-    "type": "adam",
-    "lr": 0.0001,
-    "weight_decay": 0.0
-  },
-  "lr_scheduler": {
-    "type": "step",
-    "step_size": 1000,
-    "gamma": 0.5
-  },
-  "use_cuda": true,
-  "seed": 42
+  "batch_size": 4,
+  "eval_batch_size": 4,
+  "num_loader_workers": 2,
+  "num_eval_loader_workers": 1,
+  "run_eval": true,
+  "test_delay_epochs": 1,
+  "epochs": 5,
+  "save_step": 100,
+  "print_step": 25,
+  "eval_step": 100,
+  "mixed_precision": true,
+  "cudnn_benchmark": true,
+  "lr": 1e-4,
+  "grad_clip": 5.0,
+  "loss_masking": true,
+  "use_speaker_embedding": true,
+  "compute_input_seq_cache": true,
+  "eval_split_size": 0.05,
+  "text_cleaner": "multilingual_cleaners",
+  "use_phonemes": false,
+  "phoneme_language": "es",
+  "min_text_len": 1,
+  "max_text_len": 200,
+  "min_audio_len": 1000,
+  "max_audio_len": 160000,
+  "output_step": 25,
+  "logger": "tensorboard",
+  "seed": 1234
 }