fix: resolve path issue for inference_gui on Windows

GPT_SoVITS/inference_gui.py

@@ -1,11 +1,13 @@
 import os
 import sys
+
+import soundfile as sf
 from PyQt5.QtCore import QEvent
 from PyQt5.QtWidgets import QApplication, QMainWindow, QLabel, QLineEdit, QPushButton, QTextEdit
 from PyQt5.QtWidgets import QGridLayout, QVBoxLayout, QWidget, QFileDialog, QStatusBar, QComboBox
-import soundfile as sf
 
 from tools.i18n.i18n import I18nAuto
+
 i18n = I18nAuto()
 
 from inference_webui import gpt_path, sovits_path, change_gpt_weights, change_sovits_weights, get_tts_wav
@@ -61,11 +63,11 @@ class GPTSoVITSGUI(QMainWindow):
                 border: 1px solid #45a049;
                 box-shadow: 2px 2px 2px rgba(0, 0, 0, 0.1);
             }
-        """)    
+        """)
 
         license_text = (
-        "本软件以MIT协议开源, 作者不对软件具备任何控制力, 使用软件者、传播软件导出的声音者自负全责. "
+            "本软件以MIT协议开源, 作者不对软件具备任何控制力, 使用软件者、传播软件导出的声音者自负全责. "
-        "如不认可该条款, 则不能使用或引用软件包内任何代码和文件. 详见根目录LICENSE.")
+            "如不认可该条款, 则不能使用或引用软件包内任何代码和文件. 详见根目录LICENSE.")
         license_label = QLabel(license_text)
         license_label.setWordWrap(True)
 
@@ -284,17 +286,17 @@ class GPTSoVITSGUI(QMainWindow):
             change_sovits_weights(sovits_path=SoVITS_model_path)
             self.SoVITS_Path = SoVITS_model_path
 
-        synthesis_result = get_tts_wav(ref_wav_path=ref_audio_path, 
+        synthesis_result = get_tts_wav(ref_wav_path=ref_audio_path,
-                                       prompt_text=ref_text, 
+                                       prompt_text=ref_text,
-                                       prompt_language=language_combobox, 
+                                       prompt_language=language_combobox,
-                                       text=target_text, 
+                                       text=target_text,
                                        text_language=target_language_combobox)
 
         result_list = list(synthesis_result)
 
         if result_list:
             last_sampling_rate, last_audio_data = result_list[-1]
-            output_wav_path = os.path.join(output_path, "output.wav") 
+            output_wav_path = os.path.join(output_path, "output.wav")
             sf.write(output_wav_path, last_audio_data, last_sampling_rate)
 
             result = "Audio saved to " + output_wav_path
@@ -307,4 +309,4 @@ if __name__ == '__main__':
     app = QApplication(sys.argv)
     mainWin = GPTSoVITSGUI()
     mainWin.show()
-    sys.exit(app.exec_())
+    sys.exit(app.exec_())

GPT_SoVITS/inference_webui.py

@@ -18,31 +18,50 @@ logging.getLogger("torchaudio._extension").setLevel(logging.ERROR)
 import pdb
 import torch
 
-if os.path.exists("./gweight.txt"):
+# 获取项目根目录的绝对路径
-    with open("./gweight.txt", 'r', encoding="utf-8") as file:
+PROJECT_ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
-        gweight_data = file.read()
-        gpt_path = os.environ.get(
-            "gpt_path", gweight_data)
-else:
-    gpt_path = os.environ.get(
-        "gpt_path", "GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt")
 
-if os.path.exists("./sweight.txt"):
+def get_absolute_path(relative_path):
-    with open("./sweight.txt", 'r', encoding="utf-8") as file:
+    # 如果路径已经是绝对路径，直接返回
+    if os.path.isabs(relative_path):
+        return relative_path
+    return os.path.join(PROJECT_ROOT, relative_path)
+
+# 检查和加载 sovits_path
+sweight_file_path = get_absolute_path("sweight.txt")
+if os.path.exists(sweight_file_path):
+    with open(sweight_file_path, 'r', encoding="utf-8") as file:
+        sovits_path = file.read().strip()
+        sovits_path = get_absolute_path(sovits_path)
+else:
+    sovits_path = get_absolute_path("GPT_SoVITS/pretrained_models/s2G488k.pth")
+
+# 检查和加载 gpt_path
+gweight_file_path = get_absolute_path("gweight.txt")
+if os.path.exists(gweight_file_path):
+    with open(gweight_file_path, 'r', encoding="utf-8") as file:
+        gpt_path = file.read().strip()
+        gpt_path = get_absolute_path(gpt_path)
+else:
+    gpt_path = get_absolute_path("GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt")
+
+if os.path.exists(gweight_file_path):
+    with open(gweight_file_path, 'r', encoding="utf-8") as file:
+        gweight_data = file.read()
+        gpt_path = os.environ.get("gpt_path", gweight_data)
+else:
+    gpt_path = os.environ.get("gpt_path", get_absolute_path("GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt"))
+
+if os.path.exists(sweight_file_path):
+    with open(sweight_file_path, 'r', encoding="utf-8") as file:
         sweight_data = file.read()
         sovits_path = os.environ.get("sovits_path", sweight_data)
 else:
-    sovits_path = os.environ.get("sovits_path", "GPT_SoVITS/pretrained_models/s2G488k.pth")
+    sovits_path = os.environ.get("sovits_path", get_absolute_path("GPT_SoVITS/pretrained_models/s2G488k.pth"))
-# gpt_path = os.environ.get(
+
-#     "gpt_path", "pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt"
+cnhubert_base_path = os.environ.get("cnhubert_base_path", get_absolute_path("GPT_SoVITS/pretrained_models/chinese-hubert-base"))
-# )
+bert_path = os.environ.get("bert_path", get_absolute_path("GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large"))
-# sovits_path = os.environ.get("sovits_path", "pretrained_models/s2G488k.pth")
+
-cnhubert_base_path = os.environ.get(
-    "cnhubert_base_path", "GPT_SoVITS/pretrained_models/chinese-hubert-base"
-)
-bert_path = os.environ.get(
-    "bert_path", "GPT_SoVITS/pretrained_models/chinese-roberta-wwm-ext-large"
-)
 infer_ttswebui = os.environ.get("infer_ttswebui", 9872)
 infer_ttswebui = int(infer_ttswebui)
 is_share = os.environ.get("is_share", "False")
@@ -138,6 +157,7 @@ else:
 
 def change_sovits_weights(sovits_path):
     global vq_model, hps
+    sovits_path = get_absolute_path(sovits_path)
     dict_s2 = torch.load(sovits_path, map_location="cpu")
     hps = dict_s2["config"]
     hps = DictToAttrRecursive(hps)
@@ -148,38 +168,43 @@ def change_sovits_weights(sovits_path):
         n_speakers=hps.data.n_speakers,
         **hps.model
     )
-    if ("pretrained" not in sovits_path):
+    if "pretrained" not in sovits_path:
         del vq_model.enc_q
-    if is_half == True:
+    if is_half:
         vq_model = vq_model.half().to(device)
     else:
         vq_model = vq_model.to(device)
     vq_model.eval()
     print(vq_model.load_state_dict(dict_s2["weight"], strict=False))
-    with open("./sweight.txt", "w", encoding="utf-8") as f:
+
+    sweight_file_path = get_absolute_path("sweight.txt")
+    with open(sweight_file_path, "w", encoding="utf-8") as f:
         f.write(sovits_path)
 
 
-change_sovits_weights(sovits_path)
-
-
 def change_gpt_weights(gpt_path):
     global hz, max_sec, t2s_model, config
+    gpt_path = get_absolute_path(gpt_path)
     hz = 50
     dict_s1 = torch.load(gpt_path, map_location="cpu")
     config = dict_s1["config"]
     max_sec = config["data"]["max_sec"]
     t2s_model = Text2SemanticLightningModule(config, "****", is_train=False)
     t2s_model.load_state_dict(dict_s1["weight"])
-    if is_half == True:
+    if is_half:
         t2s_model = t2s_model.half()
     t2s_model = t2s_model.to(device)
     t2s_model.eval()
     total = sum([param.nelement() for param in t2s_model.parameters()])
     print("Number of parameter: %.2fM" % (total / 1e6))
-    with open("./gweight.txt", "w", encoding="utf-8") as f: f.write(gpt_path)
+
+    gweight_file_path = get_absolute_path("gweight.txt")
+    with open(gweight_file_path, "w", encoding="utf-8") as f:
+        f.write(gpt_path)
 
 
+# 更新权重路径
+change_sovits_weights(sovits_path)
 change_gpt_weights(gpt_path)
 
 
@@ -554,10 +579,11 @@ def change_choices():
     return {"choices": sorted(SoVITS_names, key=custom_sort_key), "__type__": "update"}, {"choices": sorted(GPT_names, key=custom_sort_key), "__type__": "update"}
 
 
-pretrained_sovits_name = "GPT_SoVITS/pretrained_models/s2G488k.pth"
+pretrained_sovits_name = get_absolute_path("GPT_SoVITS/pretrained_models/s2G488k.pth")
-pretrained_gpt_name = "GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt"
+pretrained_gpt_name = get_absolute_path("GPT_SoVITS/pretrained_models/s1bert25hz-2kh-longer-epoch=68e-step=50232.ckpt")
-SoVITS_weight_root = "SoVITS_weights"
+SoVITS_weight_root = get_absolute_path("SoVITS_weights")
-GPT_weight_root = "GPT_weights"
+GPT_weight_root = get_absolute_path("GPT_weights")
+
 os.makedirs(SoVITS_weight_root, exist_ok=True)
 os.makedirs(GPT_weight_root, exist_ok=True)
 

tools/i18n/i18n.py

@@ -2,9 +2,14 @@ import json
 import locale
 import os
 
+# 获取项目根目录的绝对路径
+PROJECT_ROOT = os.path.abspath(os.path.join(os.path.dirname(__file__), '../../'))
+
 
 def load_language_list(language):
-    with open(f"./i18n/locale/{language}.json", "r", encoding="utf-8") as f:
+    # 使用从项目根目录开始的绝对路径
+    language_file_path = os.path.join(PROJECT_ROOT, f"i18n/locale/{language}.json")
+    with open(language_file_path, "r", encoding="utf-8") as f:
         language_list = json.load(f)
     return language_list
 
@@ -15,7 +20,8 @@ class I18nAuto:
             language = locale.getdefaultlocale()[
                 0
             ]  # getlocale can't identify the system's language ((None, None))
-        if not os.path.exists(f"./i18n/locale/{language}.json"):
+        language_file_path = os.path.join(PROJECT_ROOT, f"i18n/locale/{language}.json")
+        if not os.path.exists(language_file_path):
             language = "en_US"
         self.language = language
         self.language_map = load_language_list(language)