update

app.py

@@ -29,20 +29,20 @@ from TTS.tts.configs.xtts_config import XttsConfig
 from TTS.tts.models.xtts import Xtts
 from TTS.utils.generic_utils import get_user_data_dir
 
-HF_TOKEN = os.environ.get("HF_TOKEN")
+#HF_TOKEN = os.environ.get("HF_TOKEN")
 
 from huggingface_hub import HfApi
 
 # will use api to restart space on a unrecoverable error
-api = HfApi(token=HF_TOKEN)
+#api = HfApi(token=HF_TOKEN)
 repo_id = "coqui/xtts-streaming"
 
-# Use never ffmpeg binary for Ubuntu20 to use denoising for microphone input
-print("Export newer ffmpeg binary for denoise filter")
-ZipFile("ffmpeg.zip").extractall()
-print("Make ffmpeg binary executable")
-st = os.stat('ffmpeg')
-os.chmod('ffmpeg', st.st_mode | stat.S_IEXEC)
+## Use never ffmpeg binary for Ubuntu20 to use denoising for microphone input
+#print("Export newer ffmpeg binary for denoise filter")
+#ZipFile("ffmpeg.zip").extractall()
+#print("Make ffmpeg binary executable")
+#st = os.stat('ffmpeg')
+#os.chmod('ffmpeg', st.st_mode | stat.S_IEXEC)
 
 # This will trigger downloading model
 print("Downloading if not downloaded Coqui XTTS V1.1")
@@ -151,7 +151,8 @@ def predict(prompt, language, audio_file_pth, mic_file_path, use_mic, voice_clea
                 out_filename = speaker_wav + str(uuid.uuid4()) + ".wav"  #ffmpeg to know output format
     
                 #we will use newer ffmpeg as that has afftn denoise filter
-                shell_command = f"./ffmpeg -y -i {speaker_wav} -af {lowpass_highpass}{trim_silence} {out_filename}".split(" ")
+                #shell_command = f"./ffmpeg -y -i {speaker_wav} -af {lowpass_highpass}{trim_silence} {out_filename}".split(" ")
+                shell_command = f"ffmpeg -y -i {speaker_wav} -af {lowpass_highpass}{trim_silence} {out_filename}".split(" ")
     
                 command_result = subprocess.run([item for item in shell_command], capture_output=False,text=True, check=True)
                 speaker_wav=out_filename
@@ -277,7 +278,7 @@ def predict(prompt, language, audio_file_pth, mic_file_path, use_mic, voice_clea
                 )
                 
                 # HF Space specific.. This error is unrecoverable need to restart space 
-                api.restart_space(repo_id=repo_id)
+                #api.restart_space(repo_id=repo_id)
             else:
                 if "Failed to decode" in str(e):
                     print("Speaker encoding error", str(e))
@@ -560,4 +561,4 @@ gr.Interface(
     article=article,
     examples=examples,
     cache_examples=False,
-).queue().launch(debug=True,show_api=True)
+).queue().launch(debug=True,show_api=True)

requirements.txt

@@ -1,7 +1,7 @@
 # Preinstall requirements from TTS 
-torch==2.0.1 --index-url https://download.pytorch.org/whl/cu118
-torchvision==0.15.2 --index-url https://download.pytorch.org/whl/cu118 
-torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118
+#torch==2.0.1 --index-url https://download.pytorch.org/whl/cu118
+#torchvision==0.15.2 --index-url https://download.pytorch.org/whl/cu118 
+#torchaudio==2.0.2 --index-url https://download.pytorch.org/whl/cu118
 numpy==1.22.0;python_version<="3.10"
 numpy==1.24.3;python_version>"3.10"
 cython==0.29.30
@@ -59,4 +59,4 @@ deepspeed==0.8.3
 pydub
 cutlet
 mecab-python3==1.0.6
-unidic-lite==1.0.8
+unidic-lite==1.0.8