aboutsummarylogtreecommitdiffstats
path: root/README.md
blob: 759dc285083803b9e4588f8a859b4ac48b86a10c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
# archlinux build for localai git version

## fixme

- make protobuf dependency on current system version for python deps. and others
- make onnxruntime dependency on current system
- llama.cpp makes gpu reboot after ~ localai 2.16?

## build

Architectures:

- cpu
- cuda
- rocm (tested on hardware)

Integrated backends:

- Audio Transcription
    - whisper
- Files
    - local-store
- GPT, Embeddings
    - llama-cpp

Python backends:

- Audio Generation
    - bark
    - coqui-tts (OK)
    - parler-tts
- Embeddings (OK)
    - sentencetransformers
- GPT, Embeddings
    - transformers
- Image generation
    - diffusers
- Information Ranking
    - rerankers
- Musik Generation
    - transformers-musicgen

### Bark

ROCM: works, but very slow

#### Speaker Library

Available:

- https://suno-ai.notion.site/8b8e8749ed514b0cbf3f699013548683?v=bc67cff786b04b50b3ceb756fd05f68c

Tested:

- english/male:   v2/en_speaker_6
- english/female: v2/en_speaker_9
- german/male:    v2/de_speaker_0
- german/female:  v2/de_speaker_3
- german/female:  v2/de_speaker_8

### Coqui-tts

ROCM: tested works on rocm gpu, some models need espeak-ng, some dont work

Tested:

- tts_models/de/thorsten/tacotron2-DCA
- tts_models/en/ljspeech/tacotron2-DDC_ph

Available:

- tts_models/multilingual/multi-dataset/xtts_v2
- tts_models/multilingual/multi-dataset/xtts_v1.1
- tts_models/multilingual/multi-dataset/your_tts
- tts_models/multilingual/multi-dataset/bark
- tts_models/bg/cv/vits
- tts_models/cs/cv/vits
- tts_models/da/cv/vits
- tts_models/et/cv/vits
- tts_models/ga/cv/vits
- tts_models/en/ek1/tacotron2
- tts_models/en/ljspeech/tacotron2-DDC
- tts_models/en/ljspeech/tacotron2-DDC_ph
- tts_models/en/ljspeech/glow-tts
- tts_models/en/ljspeech/speedy-speech
- tts_models/en/ljspeech/tacotron2-DCA
- tts_models/en/ljspeech/vits
- tts_models/en/ljspeech/vits--neon
- tts_models/en/ljspeech/fast_pitch
- tts_models/en/ljspeech/overflow
- tts_models/en/ljspeech/neural_hmm
- tts_models/en/vctk/vits
- tts_models/en/vctk/fast_pitch
- tts_models/en/sam/tacotron-DDC
- tts_models/en/blizzard2013/capacitron-t2-c50
- tts_models/en/blizzard2013/capacitron-t2-c150_v2
- tts_models/en/multi-dataset/tortoise-v2
- tts_models/en/jenny/jenny
- tts_models/es/mai/tacotron2-DDC
- tts_models/es/css10/vits
- tts_models/fr/mai/tacotron2-DDC
- tts_models/fr/css10/vits
- tts_models/uk/mai/glow-tts
- tts_models/uk/mai/vits
- tts_models/zh-CN/baker/tacotron2-DDC-GST
- tts_models/nl/mai/tacotron2-DDC
- tts_models/nl/css10/vits
- tts_models/de/thorsten/tacotron2-DCA
- tts_models/de/thorsten/vits
- tts_models/de/thorsten/tacotron2-DDC
- tts_models/de/css10/vits-neon
- tts_models/ja/kokoro/tacotron2-DDC
- tts_models/tr/common-voice/glow-tts
- tts_models/it/mai_female/glow-tts
- tts_models/it/mai_female/vits
- tts_models/it/mai_male/glow-tts
- tts_models/it/mai_male/vits
- tts_models/ewe/openbible/vits
- tts_models/hau/openbible/vits
- tts_models/lin/openbible/vits
- tts_models/tw_akuapem/openbible/vits
- tts_models/tw_asante/openbible/vits
- tts_models/yor/openbible/vits
- tts_models/hu/css10/vits
- tts_models/el/cv/vits
- tts_models/fi/css10/vits
- tts_models/hr/cv/vits
- tts_models/lt/cv/vits
- tts_models/lv/cv/vits
- tts_models/mt/cv/vits
- tts_models/pl/mai_female/vits
- tts_models/pt/cv/vits
- tts_models/ro/cv/vits
- tts_models/sk/cv/vits
- tts_models/sl/cv/vits
- tts_models/sv/cv/vits
- tts_models/ca/custom/vits
- tts_models/fa/custom/glow-tts
- tts_models/bn/custom/vits-male
- tts_models/bn/custom/vits-female
- tts_models/be/common-voice/glow-tts
- vocoder_models/universal/libri-tts/wavegrad
- vocoder_models/universal/libri-tts/fullband-melgan
- vocoder_models/en/ek1/wavegrad
- vocoder_models/en/ljspeech/multiband-melgan
- vocoder_models/en/ljspeech/hifigan_v2
- vocoder_models/en/ljspeech/univnet
- vocoder_models/en/blizzard2013/hifigan_v2
- vocoder_models/en/vctk/hifigan_v2
- vocoder_models/en/sam/hifigan_v2
- vocoder_models/nl/mai/parallel-wavegan
- vocoder_models/de/thorsten/wavegrad
- vocoder_models/de/thorsten/fullband-melgan
- vocoder_models/de/thorsten/hifigan_v1
- vocoder_models/ja/kokoro/hifigan_v1
- vocoder_models/uk/mai/multiband-melgan
- vocoder_models/tr/common-voice/hifigan
- vocoder_models/be/common-voice/hifigan
- voice_conversion_models/multilingual/vctk/freevc24