Hugging Face's logo

BricksDisplay
/
NeuTTS-Air-GGUF

Text-to-Speech
GGUF
English
tts
neutts
Model card Files
xet
 Community
hans00 commited on
Commit
2d14e88
·
verified ·
1 Parent(s): 50ded1d

Initial NeuTTS Air GGUF release

Browse files
Files changed (24) hide show
  1. .gitattributes +22 -0
  2. README.md +74 -0
  3. codec-f16.gguf +3 -0
  4. codec-f32.gguf +3 -0
  5. codec-q4_k_m.gguf +3 -0
  6. codec-q5_k_m.gguf +3 -0
  7. codec-q8_0.gguf +3 -0
  8. llm-bf16.gguf +3 -0
  9. llm-f16.gguf +3 -0
  10. llm-f32.gguf +3 -0
  11. llm-q2_k.gguf +3 -0
  12. llm-q3_k_l.gguf +3 -0
  13. llm-q3_k_m.gguf +3 -0
  14. llm-q3_k_s.gguf +3 -0
  15. llm-q4_0.gguf +3 -0
  16. llm-q4_1.gguf +3 -0
  17. llm-q4_k_m.gguf +3 -0
  18. llm-q4_k_s.gguf +3 -0
  19. llm-q5_0.gguf +3 -0
  20. llm-q5_1.gguf +3 -0
  21. llm-q5_k_m.gguf +3 -0
  22. llm-q5_k_s.gguf +3 -0
  23. llm-q6_k.gguf +3 -0
  24. llm-q8_0.gguf +3 -0
.gitattributes CHANGED
@@ -33,3 +33,25 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ codec-f16.gguf filter=lfs diff=lfs merge=lfs -text
37
+ codec-f32.gguf filter=lfs diff=lfs merge=lfs -text
38
+ codec-q4_k_m.gguf filter=lfs diff=lfs merge=lfs -text
39
+ codec-q5_k_m.gguf filter=lfs diff=lfs merge=lfs -text
40
+ codec-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
41
+ llm-bf16.gguf filter=lfs diff=lfs merge=lfs -text
42
+ llm-f16.gguf filter=lfs diff=lfs merge=lfs -text
43
+ llm-f32.gguf filter=lfs diff=lfs merge=lfs -text
44
+ llm-q2_k.gguf filter=lfs diff=lfs merge=lfs -text
45
+ llm-q3_k_l.gguf filter=lfs diff=lfs merge=lfs -text
46
+ llm-q3_k_m.gguf filter=lfs diff=lfs merge=lfs -text
47
+ llm-q3_k_s.gguf filter=lfs diff=lfs merge=lfs -text
48
+ llm-q4_0.gguf filter=lfs diff=lfs merge=lfs -text
49
+ llm-q4_1.gguf filter=lfs diff=lfs merge=lfs -text
50
+ llm-q4_k_m.gguf filter=lfs diff=lfs merge=lfs -text
51
+ llm-q4_k_s.gguf filter=lfs diff=lfs merge=lfs -text
52
+ llm-q5_0.gguf filter=lfs diff=lfs merge=lfs -text
53
+ llm-q5_1.gguf filter=lfs diff=lfs merge=lfs -text
54
+ llm-q5_k_m.gguf filter=lfs diff=lfs merge=lfs -text
55
+ llm-q5_k_s.gguf filter=lfs diff=lfs merge=lfs -text
56
+ llm-q6_k.gguf filter=lfs diff=lfs merge=lfs -text
57
+ llm-q8_0.gguf filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,74 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: other
3
+ language:
4
+ - en
5
+ tags:
6
+ - gguf
7
+ - tts
8
+ - neutts
9
+ - text-to-speech
10
+ base_model: neuphonic/neutts-air
11
+ ---
12
+
13
+ # NeuTTS Air GGUF
14
+
15
+ GGUF conversions of [Neuphonic's NeuTTS Air](https://huggingface.co/neuphonic/neutts-air) — the larger (~750 M) on-device TTS LM in the NeuTTS family with instant voice cloning.
16
+
17
+ Split:
18
+
19
+ - **LLM-part** — text-and-reference-codes to speech token IDs, stock `qwen2` arch, runs in **llama.cpp**.
20
+ - **Codec-part** — speech token IDs to/from 24 kHz PCM via [NeuCodec](https://huggingface.co/neuphonic/neucodec), runs in [codec.cpp](https://github.com/mybigday/codec.cpp).
21
+
22
+ ## Files
23
+
24
+ ### LLM-part (qwen2 arch, vocab 217652, RoPE θ=1000000)
25
+
26
+ `llm-<quant>.gguf`
27
+
28
+ | File | Size |
29
+ | ----------------- | ------- |
30
+ | `llm-f32.gguf` | 2.99 GB |
31
+ | `llm-f16.gguf` | 1.50 GB |
32
+ | `llm-bf16.gguf` | 1.50 GB |
33
+ | `llm-q8_0.gguf` | 803 MB |
34
+ | `llm-q6_k.gguf` | 777 MB |
35
+ | `llm-q5_k_m.gguf` | 631 MB |
36
+ | `llm-q5_k_s.gguf` | 623 MB |
37
+ | `llm-q5_1.gguf` | 630 MB |
38
+ | `llm-q5_0.gguf` | 595 MB |
39
+ | `llm-q4_k_m.gguf` | 596 MB |
40
+ | `llm-q4_k_s.gguf` | 584 MB |
41
+ | `llm-q4_1.gguf` | 561 MB |
42
+ | `llm-q4_0.gguf` | 526 MB |
43
+ | `llm-q3_k_l.gguf` | 543 MB |
44
+ | `llm-q3_k_m.gguf` | 530 MB |
45
+ | `llm-q3_k_s.gguf` | 512 MB |
46
+ | `llm-q2_k.gguf` | 513 MB |
47
+
48
+ ### Codec-part (NeuCodec, `neucodec` arch, 65536 codebook, 24 kHz, 80 Hz token rate)
49
+
50
+ `codec[-<quant>].gguf`
51
+
52
+ | File | Size |
53
+ | ------------------- | ------- |
54
+ | `codec-f32.gguf` | 928 MB |
55
+ | `codec-f16.gguf` | 465 MB |
56
+ | `codec-q8_0.gguf` | 326 MB |
57
+ | `codec-q5_k_m.gguf` | 270 MB |
58
+ | `codec-q4_k_m.gguf` | 252 MB |
59
+
60
+ ## Prompt format
61
+
62
+ ```
63
+ user: Convert the text to speech:<|TEXT_PROMPT_START|>{ref_phones} {input_phones}<|TEXT_PROMPT_END|>
64
+ assistant:<|SPEECH_GENERATION_START|>{ref_codes_as_speech_tokens}
65
+ ```
66
+
67
+ Speech-token vocab range: `<|speech_0|>` = id 151671 … `<|speech_65535|>` = id 217206.
68
+ Stop token: `<|SPEECH_GENERATION_END|>` = id 151670.
69
+
70
+ Phonemization (`{*_phones}`) and reference-audio encoding into `<|speech_N|>` tokens are host-side concerns; see the NeuTTS Nano repo / upstream `neuphonic/neutts` for the reference Python flow.
71
+
72
+ ## Notes
73
+
74
+ - Source weights: `neuphonic/neutts-air` (LLM) and `neuphonic/neucodec` (codec).
codec-f16.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b4c9ba9bd58c300bff5104d28c0783c0f534af94109435ca793bc237e08ad962
3
+ size 487088128
codec-f32.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bec04aac7748c50991bd2f4d3fc95abd58f29d880e486765e00b7cdd3eb29594
3
+ size 972715616
codec-q4_k_m.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60c52aed69b4988b749b3f6827c494faa2baa4cb87af47a7cdf7bc87b9e9dfc8
3
+ size 264189056
codec-q5_k_m.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ddb958a500260cd7e36f3adaf0cb6da9201c3bd3e105b12ce3ce8fa419bebfa6
3
+ size 283571584
codec-q8_0.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9972d7cabd38582f0425ba898abffc6f92166154ce046a4f392654dbdad5fafa
3
+ size 341719168
llm-bf16.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e62462cd6ddf1a73b6bbf998a74513ecfa96ab7150ddb49fb7484e13d5af5c3f
3
+ size 1503776128
llm-f16.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f816a587501ac846f96bad22d57f3637acb3efc7a1dd7505f7a96471fd41844
3
+ size 1503776128
llm-f32.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e93ce79cbfcda570a9e14dda055e2c65b371386ef489e8730498cf4c996ebd7a
3
+ size 2999494016
llm-q2_k.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a88a87f806c291fa6a0a0f3ae811c5ae7c32e2b8af3eac20e6deeb5e4c295d24
3
+ size 512689472
llm-q3_k_l.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b07a8588f5fc5e62244f917cf3fa1dce1f3ccff74b960f80d481088b11adab01
3
+ size 543440192
llm-q3_k_m.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b62466360a4105d2c56d8209e9e04b907dc284aab120416e97f57be241550b9
3
+ size 529548608
llm-q3_k_s.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:afe13bff17fb5561565f6a34eb7e8aef572dd64af511f47492954169d8e4dab4
3
+ size 512345408
llm-q4_0.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:60ea282560586ffa19afc10d37c7ecab1be67623bd026d571bced52bb4aedcec
3
+ size 526236992
llm-q4_1.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c65cb5e29d8c089c41ce80a376d234596fe1a3dbef52defa663f732e7795be89
3
+ size 560789664
llm-q4_k_m.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c5cffd8378601dad124b2a01ce5442cb4a57e7aab74c393986978264bb5d95fb
3
+ size 596267008
llm-q4_k_s.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b207e4b07e04e09a42078ce669f4ba6970af7b133d03cb915781d546cf8bbabd
3
+ size 583930880
llm-q5_0.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4aaec83d2871fbb28943ab831b6fec981ffcf69374d19ddd4615637c80c38444
3
+ size 595342336
llm-q5_1.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0d34d32b89a9bba7d965eae3f3768ee4465d06011766250afda51bf297e0ec6e
3
+ size 629895008
llm-q5_k_m.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8f36648c0058a6ecfe9b303564ed7ecd63094d78b3dc3ef83673f8b8bcf4b050
3
+ size 630733664
llm-q5_k_s.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a73ee4c4a60bd9fe36439ca26ed30109ce81bc6865e1c016481df8c0611d2348
3
+ size 623357792
llm-q6_k.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a03d46551bf48ebf70e5f131f0a99b2d3b2bf0ed4823e148b18dc05acfcf03ab
3
+ size 777326656
llm-q8_0.gguf ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:722fdecdc562cc06ae4f5a3cf622dc8f5de904d0898758c9caeaa9b0b4c5e9e7
3
+ size 802658368