Pendrokar commited on
Commit
9ff7985
·
1 Parent(s): 57a5f00

enable MegaTTS; re-enabled llasa 3b; disabled outclassed TTS

Browse files
Files changed (1) hide show
  1. app/models.py +8 -5
app/models.py CHANGED
@@ -86,7 +86,7 @@ AVAILABLE_MODELS = {
86
  # llasa 1b TTS
87
  # 'HKUST-Audio/Llasa-1B-finetuned-for-two-speakers': 'HKUST-Audio/Llasa-1B-finetuned-for-two-speakers',
88
  # llasa 3b TTS
89
- # 'srinivasbilla/llasa-3b-tts': 'srinivasbilla/llasa-3b-tts', # ZeroGPU Pro account expired
90
  # llasa 8b TTS
91
  # 'srinivasbilla/llasa-8b-tts': 'srinivasbilla/llasa-8b-tts', # ZeroGPU Pro account expired
92
 
@@ -101,16 +101,16 @@ AVAILABLE_MODELS = {
101
  # 'Steveeeeeeen/Zonos/hybrid': 'Steveeeeeeen/Zonos',
102
 
103
  # Spark
104
- 'thunnai/SparkTTS': 'thunnai/SparkTTS',
105
 
106
  # Sesame
107
- 'sesame/csm-1b' : 'sesame/csm-1b',
108
 
109
  # Orpheus
110
  'MohamedRashad/Orpheus-TTS' : 'MohamedRashad/Orpheus-TTS',
111
 
112
  # Index TTS
113
- 'IndexTeam/IndexTTS': 'IndexTeam/IndexTTS',
114
 
115
  # Dia
116
  # 'nari-labs/Dia-1.6B': 'nari-labs/Dia-1.6B', # single speaker hallucinates
@@ -121,6 +121,9 @@ AVAILABLE_MODELS = {
121
  # OpenAudio S1 (Fish Audio)
122
  'fishaudio/openaudio-s1-mini': 'fishaudio/openaudio-s1-mini',
123
 
 
 
 
124
  # HF TTS w issues
125
  # 'fishaudio/fish-speech-1': 'fishaudio/fish-speech-1', # Discontinued for OpenAudio S1
126
  # 'LeeSangHoon/HierSpeech_TTS': 'LeeSangHoon/HierSpeech_TTS', # irresponsive to exclamation marks # 4.29
@@ -973,7 +976,7 @@ OVERRIDE_INPUTS = {
973
 
974
  # Chatterbox
975
  'ResembleAI/Chatterbox': {
976
- 'audio_prompt_path_input': DEFAULT_VOICE_SAMPLE, # voice
977
  'exaggeration_input': 0.5, # 1-2
978
  'temperature_input': 0.8, # Lower values make the output more deterministic, higher values increase randomness.
979
  'seed_num_input': 1, # Seed for random number generation, can be any integer.
 
86
  # llasa 1b TTS
87
  # 'HKUST-Audio/Llasa-1B-finetuned-for-two-speakers': 'HKUST-Audio/Llasa-1B-finetuned-for-two-speakers',
88
  # llasa 3b TTS
89
+ 'srinivasbilla/llasa-3b-tts': 'srinivasbilla/llasa-3b-tts',
90
  # llasa 8b TTS
91
  # 'srinivasbilla/llasa-8b-tts': 'srinivasbilla/llasa-8b-tts', # ZeroGPU Pro account expired
92
 
 
101
  # 'Steveeeeeeen/Zonos/hybrid': 'Steveeeeeeen/Zonos',
102
 
103
  # Spark
104
+ # 'thunnai/SparkTTS': 'thunnai/SparkTTS',
105
 
106
  # Sesame
107
+ # 'sesame/csm-1b' : 'sesame/csm-1b',
108
 
109
  # Orpheus
110
  'MohamedRashad/Orpheus-TTS' : 'MohamedRashad/Orpheus-TTS',
111
 
112
  # Index TTS
113
+ # 'IndexTeam/IndexTTS': 'IndexTeam/IndexTTS', # hallucinations on the endquotes
114
 
115
  # Dia
116
  # 'nari-labs/Dia-1.6B': 'nari-labs/Dia-1.6B', # single speaker hallucinates
 
121
  # OpenAudio S1 (Fish Audio)
122
  'fishaudio/openaudio-s1-mini': 'fishaudio/openaudio-s1-mini',
123
 
124
+ # MegaTTS
125
+ 'ByteDance/MegaTTS3': 'ByteDance/MegaTTS3',
126
+
127
  # HF TTS w issues
128
  # 'fishaudio/fish-speech-1': 'fishaudio/fish-speech-1', # Discontinued for OpenAudio S1
129
  # 'LeeSangHoon/HierSpeech_TTS': 'LeeSangHoon/HierSpeech_TTS', # irresponsive to exclamation marks # 4.29
 
976
 
977
  # Chatterbox
978
  'ResembleAI/Chatterbox': {
979
+ 'audio_prompt_path_input': handle_file('https://cdn-uploads.huggingface.co/production/uploads/642c0b71eb6e214d4f8897a3/bJvzvKbiIAZ3lX0MJDU0r.wav'), # voice; chosen by Manmay of Resemble AI org - https://huggingface.co/spaces/ResembleAI/Chatterbox/discussions/14#686cd36e9479e00d8d3fc079
980
  'exaggeration_input': 0.5, # 1-2
981
  'temperature_input': 0.8, # Lower values make the output more deterministic, higher values increase randomness.
982
  'seed_num_input': 1, # Seed for random number generation, can be any integer.