Update finetune_xtts_hf.py

finetune_xtts_hf.py

@@ -25,7 +25,7 @@ subprocess.run([sys.executable, "-m", "pip", "uninstall", "-y", "hf_transfer"])
 HF_MODEL_ID = "sob111/xtts-v2-finetuned"   # <--- cambia con tu repo en HF
 HF_TOKEN = os.environ.get("HF_TOKEN")          # Debe estar definido en tu Space/entorno
 DATASET_PATH = "/tmp/dataset"        # Ruta a tu dataset
-VOXPOPULI_PATH = "/tmp/dataset/voxpopuli_es_500"        # Ruta a tu dataset
+VOXPOPULI_PATH = "/tmp/dataset/voxpopuli_es_500_vctk"        # Ruta a tu dataset
 OUTPUT_PATH = "/tmp/output_model"
 BASE_MODEL = "coqui/XTTS-v2"
 
@@ -88,7 +88,7 @@ def extract_zip(zip_file_path, destination_path):
         print(f"❌ An unexpected error occurred: {e}")
 
 # Example usage:
-zip_file = "/home/user/app/voxpopuli_es_500.zip"
+zip_file = "/home/user/app/voxpopuli_es_500_vctk.zip"
 
 # To protect against security vulnerabilities, it is important to sanitize the destination path.
 # This prevents an attacker from using a malicious ZIP file to write outside the destination folder.
@@ -103,29 +103,29 @@ print(f"safe destination {safe_destination}")
 import json
 
 # === Convertir metadata.json → metadata.csv ===
-json_path = os.path.join(VOXPOPULI_PATH, "metadata.json")
-print(f"ruta de json {json_path}")
-csv_path = os.path.join(VOXPOPULI_PATH, "metadata.csv")
+#json_path = os.path.join(VOXPOPULI_PATH, "metadata.json")
+#print(f"ruta de json {json_path}")
+#csv_path = os.path.join(VOXPOPULI_PATH, "metadata.csv")
 
-if os.path.exists(json_path):
-    print("🔄 Convirtiendo metadata.json → metadata.csv...")
-    with open(json_path, "r", encoding="utf-8") as f:
-        data = json.load(f)
+#if os.path.exists(json_path):
+#    print("🔄 Convirtiendo metadata.json → metadata.csv...")
+#    with open(json_path, "r", encoding="utf-8") as f:
+#        data = json.load(f)
 
-    with open(csv_path, "w", encoding="utf-8", newline="") as f:
-        writer = csv.writer(f, delimiter=",", quoting=csv.QUOTE_MINIMAL)
+#    with open(csv_path, "w", encoding="utf-8", newline="") as f:
+#        writer = csv.writer(f, delimiter=",", quoting=csv.QUOTE_MINIMAL)
         
-        for entry in data:
-            path = entry["audio_filepath"]
+#        for entry in data:
+#            path = entry["audio_filepath"]
             # Quitar prefijo "voxpopuli_es_500/" si existe
-            if path.startswith("voxpopuli_es_500/"):
-                path = path.replace("voxpopuli_es_500/", "", 1)
-            text = entry["text"].replace("\n", " ").strip()
-            speaker = entry.get("speaker", "spk1")
-            writer.writerow([path, text, speaker])
-    print(f"✅ metadata.csv generado en {csv_path}")
-else:
-    raise FileNotFoundError(f"❌ No se encontró {json_path}. Verifica el zip.")
+#            if path.startswith("voxpopuli_es_500/"):
+#                path = path.replace("voxpopuli_es_500/", "", 1)
+#            text = entry["text"].replace("\n", " ").strip()
+#            speaker = entry.get("speaker", "spk1")
+#            writer.writerow([path, text, speaker])
+#    print(f"✅ metadata.csv generado en {csv_path}")
+#else:
+#    raise FileNotFoundError(f"❌ No se encontró {json_path}. Verifica el zip.")
 
 from TTS.tts.datasets import load_tts_samples
 from TTS.config.shared_configs import BaseDatasetConfig
@@ -134,30 +134,28 @@ from TTS.config.shared_configs import BaseDatasetConfig
 config_dataset = BaseDatasetConfig(
     formatter="vctk",
     dataset_name="voxpopuli",
-    path="/tmp/dataset/voxpopuli_es_500",
-    meta_file_train="metadata.csv",
-    meta_file_val="metadata.csv",
+    path="/tmp/dataset/voxpopuli_es_500_vctk",
     language="es",
 )
 
 # Construimos rutas completas
-root_path = config_dataset.path
-meta_file_train = config_dataset.meta_file_train
+#root_path = config_dataset.path
+#meta_file_train = config_dataset.meta_file_train
 
-meta_path = os.path.join(root_path, meta_file_train)
-print(f"Verificando archivo CSV: {meta_path}")
-print(f"Existe?: {os.path.exists(meta_path)}")
+#meta_path = os.path.join(root_path, meta_file_train)
+#print(f"Verificando archivo CSV: {meta_path}")
+#print(f"Existe?: {os.path.exists(meta_path)}")
 
 # Intentamos cargar los samples
-try:
-    train_samples, eval_samples = load_tts_samples(config_dataset)
+#try:
+#    train_samples, eval_samples = load_tts_samples(config_dataset)
     
-    print(f"Samples detectados: {len(train_samples)} training, {len(eval_samples)} eval")
-    print("Primeros 3 samples:")
-    for s in train_samples[:3]:
-        print(s)
-except AssertionError as e:
-    print("❌ Error cargando samples:", e)
+#    print(f"Samples detectados: {len(train_samples)} training, {len(eval_samples)} eval")
+#    print("Primeros 3 samples:")
+#    for s in train_samples[:3]:
+#        print(s)
+#except AssertionError as e:
+#    print("❌ Error cargando samples:", e)
 
 print("=== Editando configuración para fine-tuning con VoxPopuli ===")