jbetker commited on
Commit
2ca4ea9
1 Parent(s): 29b2f36

Update notebook

Browse files
Files changed (1) hide show
  1. tortoise_tts.ipynb +67 -37
tortoise_tts.ipynb CHANGED
@@ -40,7 +40,8 @@
40
  "source": [
41
  "!git clone https://github.com/neonbjb/tortoise-tts.git\n",
42
  "%cd tortoise-tts\n",
43
- "!pip install -r requirements.txt"
 
44
  ]
45
  },
46
  {
@@ -54,8 +55,8 @@
54
  "\n",
55
  "import IPython\n",
56
  "\n",
57
- "from api import TextToSpeech\n",
58
- "from utils.audio import load_audio, get_voices\n",
59
  "\n",
60
  "# This will download all the models used by Tortoise from the HF hub.\n",
61
  "tts = TextToSpeech()"
@@ -66,20 +67,6 @@
66
  "execution_count": null,
67
  "outputs": []
68
  },
69
- {
70
- "cell_type": "code",
71
- "source": [
72
- "# List all the voices available. These are just some random clips I've gathered\n",
73
- "# from the internet as well as a few voices from the training dataset.\n",
74
- "# Feel free to add your own clips to the voices/ folder.\n",
75
- "%ls voices"
76
- ],
77
- "metadata": {
78
- "id": "SSleVnRAiEE2"
79
- },
80
- "execution_count": null,
81
- "outputs": []
82
- },
83
  {
84
  "cell_type": "code",
85
  "source": [
@@ -94,8 +81,6 @@
94
  "Though as for that the passing there\n",
95
  "Had worn them really about the same,\"\"\"\n",
96
  "\n",
97
- "# Pick one of the voices from above\n",
98
- "voice = 'train_dotrice'\n",
99
  "# Pick a \"preset mode\" to determine quality. Options: {\"ultra_fast\", \"fast\" (default), \"standard\", \"high_quality\"}. See docs in api.py\n",
100
  "preset = \"fast\""
101
  ],
@@ -108,15 +93,32 @@
108
  {
109
  "cell_type": "code",
110
  "source": [
111
- "# Fetch the voice references and forward execute!\n",
112
- "voices = get_voices()\n",
113
- "cond_paths = voices[voice]\n",
114
- "conds = []\n",
115
- "for cond_path in cond_paths:\n",
116
- " c = load_audio(cond_path, 22050)\n",
117
- " conds.append(c)\n",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
118
  "\n",
119
- "gen = tts.tts_with_preset(text, conds, preset)\n",
 
 
 
120
  "torchaudio.save('generated.wav', gen.squeeze(0).cpu(), 24000)\n",
121
  "IPython.display.Audio('generated.wav')"
122
  ],
@@ -129,19 +131,29 @@
129
  {
130
  "cell_type": "code",
131
  "source": [
132
- "# You can add as many conditioning voices as you want together. Combining\n",
133
- "# clips from multiple voices takes the mean of the latent space for all\n",
134
- "# voices. This creates a novel voice that is a combination of the two inputs.\n",
 
 
 
 
 
 
 
 
 
 
 
 
 
 
135
  "#\n",
136
  "# Lets see what it would sound like if Picard and Kirk had a kid with a penchant for philosophy:\n",
137
- "conds = []\n",
138
- "for v in ['pat', 'william']:\n",
139
- " cond_paths = voices[v]\n",
140
- " for cond_path in cond_paths:\n",
141
- " c = load_audio(cond_path, 22050)\n",
142
- " conds.append(c)\n",
143
  "\n",
144
- "gen = tts.tts_with_preset(\"They used to say that if man was meant to fly, he’d have wings. But he did fly. He discovered he had to.\", conds, preset)\n",
 
145
  "torchaudio.save('captain_kirkard.wav', gen.squeeze(0).cpu(), 24000)\n",
146
  "IPython.display.Audio('captain_kirkard.wav')"
147
  ],
@@ -150,6 +162,24 @@
150
  },
151
  "execution_count": null,
152
  "outputs": []
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
153
  }
154
  ]
155
  }
 
40
  "source": [
41
  "!git clone https://github.com/neonbjb/tortoise-tts.git\n",
42
  "%cd tortoise-tts\n",
43
+ "!pip3 install -r requirements.txt\n",
44
+ "!python3 setup.py install"
45
  ]
46
  },
47
  {
 
55
  "\n",
56
  "import IPython\n",
57
  "\n",
58
+ "from tortoise.api import TextToSpeech\n",
59
+ "from tortoise.utils.audio import load_audio, load_voice, load_voices\n",
60
  "\n",
61
  "# This will download all the models used by Tortoise from the HF hub.\n",
62
  "tts = TextToSpeech()"
 
67
  "execution_count": null,
68
  "outputs": []
69
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
70
  {
71
  "cell_type": "code",
72
  "source": [
 
81
  "Though as for that the passing there\n",
82
  "Had worn them really about the same,\"\"\"\n",
83
  "\n",
 
 
84
  "# Pick a \"preset mode\" to determine quality. Options: {\"ultra_fast\", \"fast\" (default), \"standard\", \"high_quality\"}. See docs in api.py\n",
85
  "preset = \"fast\""
86
  ],
 
93
  {
94
  "cell_type": "code",
95
  "source": [
96
+ "# Tortoise will attempt to mimic voices you provide. It comes pre-packaged\n",
97
+ "# with some voices you might recognize.\n",
98
+ "\n",
99
+ "# Let's list all the voices available. These are just some random clips I've gathered\n",
100
+ "# from the internet as well as a few voices from the training dataset.\n",
101
+ "# Feel free to add your own clips to the voices/ folder.\n",
102
+ "%ls tortoise/voices\n",
103
+ "\n",
104
+ "IPython.display.Audio('tortoise/voices/tom/1.wav')"
105
+ ],
106
+ "metadata": {
107
+ "id": "SSleVnRAiEE2"
108
+ },
109
+ "execution_count": null,
110
+ "outputs": []
111
+ },
112
+ {
113
+ "cell_type": "code",
114
+ "source": [
115
+ "# Pick one of the voices from the output above\n",
116
+ "voice = 'tom'\n",
117
  "\n",
118
+ "# Load it and send it through Tortoise.\n",
119
+ "voice_samples, conditioning_latents = load_voice(voice)\n",
120
+ "gen = tts.tts_with_preset(text, voice_samples=voice_samples, conditioning_latents=conditioning_latents, \n",
121
+ " preset=preset)\n",
122
  "torchaudio.save('generated.wav', gen.squeeze(0).cpu(), 24000)\n",
123
  "IPython.display.Audio('generated.wav')"
124
  ],
 
131
  {
132
  "cell_type": "code",
133
  "source": [
134
+ "# Tortoise can also generate speech using a random voice. The voice changes each time you execute this!\n",
135
+ "# (Note: random voices can be prone to strange utterances)\n",
136
+ "gen = tts.tts_with_preset(text, voice_samples=None, conditioning_latents=None, preset=preset)\n",
137
+ "torchaudio.save('generated.wav', gen.squeeze(0).cpu(), 24000)\n",
138
+ "IPython.display.Audio('generated.wav')"
139
+ ],
140
+ "metadata": {
141
+ "id": "16Xs2SSC3BXa"
142
+ },
143
+ "execution_count": null,
144
+ "outputs": []
145
+ },
146
+ {
147
+ "cell_type": "code",
148
+ "source": [
149
+ "# You can also combine conditioning voices. Combining voices produces a new voice\n",
150
+ "# with traits from all the parents.\n",
151
  "#\n",
152
  "# Lets see what it would sound like if Picard and Kirk had a kid with a penchant for philosophy:\n",
153
+ "voice_samples, conditioning_latents = load_voices(['pat', 'william'])\n",
 
 
 
 
 
154
  "\n",
155
+ "gen = tts.tts_with_preset(\"They used to say that if man was meant to fly, he’d have wings. But he did fly. He discovered he had to.\", \n",
156
+ " voice_samples=None, conditioning_latents=None, preset=preset)\n",
157
  "torchaudio.save('captain_kirkard.wav', gen.squeeze(0).cpu(), 24000)\n",
158
  "IPython.display.Audio('captain_kirkard.wav')"
159
  ],
 
162
  },
163
  "execution_count": null,
164
  "outputs": []
165
+ },
166
+ {
167
+ "cell_type": "code",
168
+ "source": [
169
+ "del tts # Will break other cells, but necessary to conserve RAM if you want to run this cell.\n",
170
+ "\n",
171
+ "# Tortoise comes with some scripts that does a lot of the lifting for you. For example,\n",
172
+ "# read.py will read a text file for you.\n",
173
+ "!python3 tortoise/read.py --voice=train_atkins --textfile=tortoise/data/riding_hood.txt --preset=ultra_fast --output_path=.\n",
174
+ "\n",
175
+ "IPython.display.Audio('train_atkins/combined.wav')\n",
176
+ "# This will take awhile.."
177
+ ],
178
+ "metadata": {
179
+ "id": "t66yqWgu68KL"
180
+ },
181
+ "execution_count": null,
182
+ "outputs": []
183
  }
184
  ]
185
  }