spybyscript commited on
Commit
6e3fd46
·
verified ·
1 Parent(s): 16f3dd0

Add Core AI and Core ML bundles for v3-20261104; model card with Apple bundles and v3-20261104 performance

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. README.md +34 -3
  2. SHA256SUMS.json +286 -2
  3. coreai/assets/fixed_references.json +0 -0
  4. coreai/assets/g2p/export_config.json +33 -0
  5. coreai/assets/g2p/language_map.json +47 -0
  6. coreai/assets/g2p/normalization.json +40 -0
  7. coreai/assets/g2p/phoneme_dict.json +9 -0
  8. coreai/assets/g2p/pronunciation_overrides.json +4 -0
  9. coreai/assets/g2p/tokenizer.json +517 -0
  10. coreai/assets/languages.json +30 -0
  11. coreai/assets/phone_vocab.json +111 -0
  12. coreai/assets/voices.json +58 -0
  13. coreai/coreai/duration_predictor.aimodel/main.hash +1 -0
  14. coreai/coreai/duration_predictor.aimodel/main.mlirb +3 -0
  15. coreai/coreai/duration_predictor.aimodel/metadata.json +5 -0
  16. coreai/coreai/g2p.aimodel/main.hash +1 -0
  17. coreai/coreai/g2p.aimodel/main.mlirb +3 -0
  18. coreai/coreai/g2p.aimodel/metadata.json +5 -0
  19. coreai/coreai/pitch_context_encoder.aimodel/main.hash +1 -0
  20. coreai/coreai/pitch_context_encoder.aimodel/main.mlirb +3 -0
  21. coreai/coreai/pitch_context_encoder.aimodel/metadata.json +5 -0
  22. coreai/coreai/pitch_estimator.aimodel/main.hash +1 -0
  23. coreai/coreai/pitch_estimator.aimodel/main.mlirb +3 -0
  24. coreai/coreai/pitch_estimator.aimodel/metadata.json +5 -0
  25. coreai/coreai/vector_context_encoder.aimodel/main.hash +1 -0
  26. coreai/coreai/vector_context_encoder.aimodel/main.mlirb +3 -0
  27. coreai/coreai/vector_context_encoder.aimodel/metadata.json +5 -0
  28. coreai/coreai/vector_estimator.aimodel/main.hash +2 -0
  29. coreai/coreai/vector_estimator.aimodel/main.mlirb +3 -0
  30. coreai/coreai/vector_estimator.aimodel/metadata.json +5 -0
  31. coreai/coreai/vocoder.aimodel/main.hash +1 -0
  32. coreai/coreai/vocoder.aimodel/main.mlirb +3 -0
  33. coreai/coreai/vocoder.aimodel/metadata.json +5 -0
  34. coreai/manifest.json +1483 -0
  35. coreml/assets/fixed_references.json +0 -0
  36. coreml/assets/g2p/export_config.json +33 -0
  37. coreml/assets/g2p/language_map.json +47 -0
  38. coreml/assets/g2p/normalization.json +40 -0
  39. coreml/assets/g2p/phoneme_dict.json +9 -0
  40. coreml/assets/g2p/pronunciation_overrides.json +4 -0
  41. coreml/assets/g2p/tokenizer.json +517 -0
  42. coreml/assets/languages.json +30 -0
  43. coreml/assets/phone_vocab.json +111 -0
  44. coreml/assets/voices.json +58 -0
  45. coreml/coreml/duration_predictor.mlmodelc/analytics/coremldata.bin +3 -0
  46. coreml/coreml/duration_predictor.mlmodelc/coremldata.bin +3 -0
  47. coreml/coreml/duration_predictor.mlmodelc/model.mil +0 -0
  48. coreml/coreml/duration_predictor.mlmodelc/weights/weight.bin +3 -0
  49. coreml/coreml/g2p.mlmodelc/analytics/coremldata.bin +3 -0
  50. coreml/coreml/g2p.mlmodelc/coremldata.bin +3 -0
README.md CHANGED
@@ -3,12 +3,12 @@ license: apache-2.0
3
  language: [en, es, it, fr, de, vi]
4
  library_name: litert
5
  pipeline_tag: text-to-speech
6
- tags: [text-to-speech, tts, litert, onnx, pytorch, multilingual, expressive-tts, duration-flow]
7
  ---
8
 
9
  # Scylla's Band v3
10
 
11
- Release **`v3-20261104`**. The [Scylla's Band runtime](https://github.com/lowkeytea/scyllasband) supports this release from version 3.0.0. The manifest declares graph contract `scyllasband_measured_controls_v3`; runtimes written for v2 releases cannot run these graphs and keep downloading their v2 tags.
12
 
13
  This release follows `v3-20261103` with a distilled pitch predictor: it makes 4 network passes per passage instead of 16, and the other models are unchanged.
14
 
@@ -24,7 +24,7 @@ python -m scyllasband speak --voice ariadne --language en_us --intensity 1 --pac
24
  -o hello.wav "It is a pleasure to meet you."
25
  ```
26
 
27
- `--flavor litert` or `--flavor onnx` chooses the bundle, installed under `scyllasband/models/<bundle>`. Downloads pin a release tag.
28
 
29
  ## Controls
30
 
@@ -68,6 +68,34 @@ The flow and the pitch predictor are distilled for few-step sampling: each was t
68
 
69
  Long text is spoken as passages within the trained target lengths (64 to 420 latent frames, about 1.4 to 9 seconds): a short first passage so audio starts quickly, then passages as long as the text allows, cut at a sentence end, else a clause, else between words, and measured with the requested delivery. Each passage sees the surrounding text as context and continues from the acoustics of the one before it, and consecutive passages are decoded as one continuous waveform. Durations are predicted deterministically; different controls can alter them.
70
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
71
  ## Voices and language coverage
72
 
73
  Fourteen voices: Ariadne, Felix, Gwen, Ink, Kincade, Max, Orpheus, Poppy, Priya, Rex, Scylla, Stone, Thistle, Tuesday.
@@ -76,6 +104,8 @@ All fourteen have Spanish (`es`), Italian (`it`), French (`fr`), German (`de`) a
76
 
77
  ## Artifacts
78
 
 
 
79
  - `litert/` (229 MB): LiteRT bundle for Linux, Windows, Android and Intel Macs. Dynamic-range INT8 weights on the fully connected, convolution and embedding layers; the vocoder's depthwise convolutions stay in full precision, which keeps them on XNNPACK. Each size is a signature of one model file, sharing weights: latent buckets every 64 frames up to 512 (then 768 and 1024), flow graphs without the 96 prefix frames for a request's first passage (up to 448 frames), and narrower duration, span-context and pitch-context (128 and 256 phones) and G2P (32–256 tokens) widths beside the full ones.
80
  - `onnx/` (291 MB): ONNX Runtime bundle with dynamic INT8 weights in the duration, context, pitch and flow transformers; the vocoder stays in full precision. Latent buckets of 128, 256, 384, 512, 768 and 1024 frames share one weight file.
81
  - `webgpu/` (291 MB): ONNX Runtime bundle for ONNX Runtime Web's WebGPU backend. Transformer weights are INT8 with float compute (`MatMulNBits`), which runs entirely on the GPU; the vocoder stays in full precision.
@@ -94,6 +124,7 @@ Release validation separates graph accuracy, frontend/timing behaviour and perce
94
  - The distilled pitch predictor at 4 Euler steps compared with the undistilled predictor at 8 Heun steps (release `v3-20261103`) by blind listening on paired lines across the voices and languages, and on the same 168-clip sweep: ASR word error rate 0.062 for both, and a MOS-predictor score 0.12 higher.
95
  - Quantized bundles compared with full precision through the runtime on lines covering every voice, each control at its limits and each language. The WebGPU bundle reproduces the full-precision take (identical frame counts on 15 of 17 lines, 0.7 dB median mel-cepstral distance). Dynamic INT8 moves some phone durations by a frame, which redraws a passage's noise and gives a different take, so the ONNX and LiteRT bundles are judged on the same sweep: ASR word error rate 0.063 (ONNX) and 0.062 (LiteRT) against 0.062 for full precision, and a MOS-predictor score within 0.02 of full precision for both.
96
  - Every LiteRT graph gives identical output with and without XNNPACK's weight cache.
 
97
 
98
  Listening is decisive; automatic scores support it. Training audio is not distributed.
99
 
 
3
  language: [en, es, it, fr, de, vi]
4
  library_name: litert
5
  pipeline_tag: text-to-speech
6
+ tags: [text-to-speech, tts, litert, onnx, coreml, coreai, pytorch, multilingual, expressive-tts, duration-flow]
7
  ---
8
 
9
  # Scylla's Band v3
10
 
11
+ Release **`v3-20261104`**. The [Scylla's Band runtime](https://github.com/lowkeytea/scyllasband) supports this release from version 3.0.0. Tag `v3-20261104-apple` adds Core ML and Core AI bundles of the same model; the runtime downloads from `v3-20261104-apple`. The manifest declares graph contract `scyllasband_measured_controls_v3`; runtimes written for v2 releases cannot run these graphs and keep downloading their v2 tags.
12
 
13
  This release follows `v3-20261103` with a distilled pitch predictor: it makes 4 network passes per passage instead of 16, and the other models are unchanged.
14
 
 
24
  -o hello.wav "It is a pleasure to meet you."
25
  ```
26
 
27
+ The default download is the Core ML bundle on an Apple silicon Mac and the LiteRT bundle elsewhere, installed under `scyllasband/models/<bundle>`. A Mac that can build apps for iOS, iPadOS or visionOS 27 is also offered the Core AI bundle those apps use. `--flavor coreml`, `--flavor coreai`, `--flavor litert` or `--flavor onnx` choose explicitly. Downloads pin a release tag.
28
 
29
  ## Controls
30
 
 
68
 
69
  Long text is spoken as passages within the trained target lengths (64 to 420 latent frames, about 1.4 to 9 seconds): a short first passage so audio starts quickly, then passages as long as the text allows, cut at a sentence end, else a clause, else between words, and measured with the requested delivery. Each passage sees the surrounding text as context and continues from the acoustics of the one before it, and consecutive passages are decoded as one continuous waveform. Durations are predicted deterministically; different controls can alter them.
70
 
71
+ ## Performance
72
+
73
+ Measured with the default samplers (flow and pitch predictor: Euler, 4 steps each) through the native runtime: `scyllasband_speak` from [libscyllasband](https://github.com/lowkeytea/scyllasband/tree/main/libscyllasband) on the Mac and Android, and the [iOS sample](https://github.com/lowkeytea/scyllasband/tree/main/examples/ios)'s benchmark mode on the iPhone. The text is three sentences, 10.3 s of audio: "The last train leaves at midnight, so pack light. Bring a warm coat, and don't forget the map. If we miss it, we walk, and nobody wants to walk that far in the rain." (voice Scylla, seed 7, the voice's own delivery).
74
+
75
+ - **First run**: from launch to first audio on the first launch after installing the bundle. Core ML compiles and Core AI specializes the graphs for the device then (Core AI caches them until the next OS update); LiteRT writes its packed weights to a cache file beside each graph.
76
+ - **Cold start**: from launch to first audio on a later launch, including loading the bundle.
77
+ - **Warm first audio**: from a request to its first audio in a process that has already spoken (median of three requests).
78
+ - **Real-time speed**: seconds of audio generated per second when warm, rounded.
79
+
80
+ | Device | Bundle | Compute | First run | Cold start | Warm first audio | Real-time speed |
81
+ | --- | --- | --- | --- | --- | --- | --- |
82
+ | MacBook Pro, M5 Max (macOS 27) | Core ML | GPU (default) | 2.73 s | 1.34 s | 30 ms | ~170× |
83
+ | | Core ML | CPU | 2.17 s | 0.33 s | 221 ms | ~20× |
84
+ | | Core AI | GPU (default) | 5.80 s | 0.46 s | 32 ms | ~150× |
85
+ | | Core AI | Neural Engine | 34.2 s | 0.60 s | 316 ms | ~20× |
86
+ | | Core AI | CPU | 3.54 s | 0.93 s | 844 ms | ~5.0× |
87
+ | | LiteRT | CPU | 0.41 s | 0.30 s | 105 ms | ~34× |
88
+ | | ONNX Runtime | CPU | 0.88 s | 0.85 s | 216 ms | ~21× |
89
+ | iPhone 17 Pro Max (iOS 27.2) | Core ML | GPU (default) | 15.5 s | 1.79 s | 77 ms | ~65× |
90
+ | | Core ML | CPU | 2.87 s | 0.49 s | 294 ms | ~15× |
91
+ | | Core AI | GPU (default) | 18.5 s | 0.59 s | 70 ms | ~68× |
92
+ | | Core AI | Neural Engine | 8.00 s | 0.74 s | 367 ms | ~17× |
93
+ | | Core AI | CPU | 7.58 s | 1.16 s | 1010 ms | ~4.1× |
94
+ | Galaxy Z Fold 8, Snapdragon SM8850 (Android 17) | LiteRT | CPU | 0.81 s | 0.52 s | 120 ms | ~27× |
95
+ | | ONNX Runtime | CPU | 1.85 s | 1.89 s | 461 ms | ~9.2× |
96
+
97
+ The GPU is the default on Apple platforms. The Neural Engine placement runs the flow and the pitch predictor on the Neural Engine and the other graphs on the CPU, for apps that speak in the background, where iOS does not allow GPU work. Core ML reloads its graphs on every launch, which accounts for most of its cold start on the GPU. Each passage runs the pitch predictor 4 times, then the flow 4 times. Against `v3-20261103`, whose pitch predictor ran 16 times, warm first audio is 30–45 % shorter on every device and bundle (12–13 % with the Neural Engine placement).
98
+
99
  ## Voices and language coverage
100
 
101
  Fourteen voices: Ariadne, Felix, Gwen, Ink, Kincade, Max, Orpheus, Poppy, Priya, Rex, Scylla, Stone, Thistle, Tuesday.
 
104
 
105
  ## Artifacts
106
 
107
+ - `coreai/` (178 MB): Core AI bundle for apps on iOS, iPadOS and visionOS 27 (`.aimodel` assets); it also runs on macOS 27, where the runtime uses the Core ML bundle by default. INT8 weights per output channel with FP16 compute; latent buckets of 128, 256, 384, 512, 768 and 1024 frames are functions of one asset sharing weights. The flow and the pitch predictor also run accurately on the Neural Engine (`controls.coreai.neural_engine_assets`), which is the placement for apps that speak in the background; the GPU is the default.
108
+ - `coreml/` (163 MB): Core ML bundle for Apple silicon with iOS 18 / visionOS 2 / macOS 15 and later (compiled `.mlmodelc` assets), with INT8 weights and FP16 compute; GPU by default, CPU in background apps.
109
  - `litert/` (229 MB): LiteRT bundle for Linux, Windows, Android and Intel Macs. Dynamic-range INT8 weights on the fully connected, convolution and embedding layers; the vocoder's depthwise convolutions stay in full precision, which keeps them on XNNPACK. Each size is a signature of one model file, sharing weights: latent buckets every 64 frames up to 512 (then 768 and 1024), flow graphs without the 96 prefix frames for a request's first passage (up to 448 frames), and narrower duration, span-context and pitch-context (128 and 256 phones) and G2P (32–256 tokens) widths beside the full ones.
110
  - `onnx/` (291 MB): ONNX Runtime bundle with dynamic INT8 weights in the duration, context, pitch and flow transformers; the vocoder stays in full precision. Latent buckets of 128, 256, 384, 512, 768 and 1024 frames share one weight file.
111
  - `webgpu/` (291 MB): ONNX Runtime bundle for ONNX Runtime Web's WebGPU backend. Transformer weights are INT8 with float compute (`MatMulNBits`), which runs entirely on the GPU; the vocoder stays in full precision.
 
124
  - The distilled pitch predictor at 4 Euler steps compared with the undistilled predictor at 8 Heun steps (release `v3-20261103`) by blind listening on paired lines across the voices and languages, and on the same 168-clip sweep: ASR word error rate 0.062 for both, and a MOS-predictor score 0.12 higher.
125
  - Quantized bundles compared with full precision through the runtime on lines covering every voice, each control at its limits and each language. The WebGPU bundle reproduces the full-precision take (identical frame counts on 15 of 17 lines, 0.7 dB median mel-cepstral distance). Dynamic INT8 moves some phone durations by a frame, which redraws a passage's noise and gives a different take, so the ONNX and LiteRT bundles are judged on the same sweep: ASR word error rate 0.063 (ONNX) and 0.062 (LiteRT) against 0.062 for full precision, and a MOS-predictor score within 0.02 of full precision for both.
126
  - Every LiteRT graph gives identical output with and without XNNPACK's weight cache.
127
+ - Core ML and Core AI bundles compared with the PyTorch model on every graph and through the runtime on 14 passages (every voice, each control at both limits, six languages, narration), on each compute unit they are used with (GPU, CPU, and for Core AI the Neural Engine): identical phonemes, fewer changed phone durations than the LiteRT bundle (0.2–0.7 % against 1.9 %), and smaller pitch-contour (0.016–0.045 semitones RMS against 0.051), latent and spectral differences. The Apple exports rewrite three computations into equivalent forms that FP16 and the Neural Engine evaluate exactly (span pooling, the modifier bit decoding, the duration model's combined phone/segment index), and scale the residual stream of the flow and the pitch predictor; in FP32 the rewritten graphs match the trained ones (identical outputs; the span encoders within 1e-6).
128
 
129
  Listening is decisive; automatic scores support it. Training audio is not distributed.
130
 
SHA256SUMS.json CHANGED
@@ -4,13 +4,297 @@
4
  "bytes": 11357
5
  },
6
  "README.md": {
7
- "sha256": "e5dc732ed5a1fc280ea50e6f185915c547f2cef7c8029000aaf01e7daac2aa06",
8
- "bytes": 9634
9
  },
10
  "THIRD_PARTY_NOTICES.md": {
11
  "sha256": "757a3eca5f8284352d471887af2c31b65c5d23d2a642f0ae77e6a11c150dba39",
12
  "bytes": 2829
13
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
14
  "litert/assets/fixed_references.json": {
15
  "sha256": "f227c55cab5f6b8cffa33c3dd743e69b26c7e446be5b9cd9a286095cabaa249e",
16
  "bytes": 1555673
 
4
  "bytes": 11357
5
  },
6
  "README.md": {
7
+ "sha256": "17a4a41f512fdc0c07a04779dae2e763ce7a05ac4aa645b213ce898434132d1a",
8
+ "bytes": 14515
9
  },
10
  "THIRD_PARTY_NOTICES.md": {
11
  "sha256": "757a3eca5f8284352d471887af2c31b65c5d23d2a642f0ae77e6a11c150dba39",
12
  "bytes": 2829
13
  },
14
+ "coreai/assets/fixed_references.json": {
15
+ "sha256": "f227c55cab5f6b8cffa33c3dd743e69b26c7e446be5b9cd9a286095cabaa249e",
16
+ "bytes": 1555673
17
+ },
18
+ "coreai/assets/g2p/export_config.json": {
19
+ "sha256": "0ec5ea81f346521f3a5c9c4a90b939c6b5143b1a11cc5e74b9902969db56e568",
20
+ "bytes": 928
21
+ },
22
+ "coreai/assets/g2p/language_map.json": {
23
+ "sha256": "b3db2048b492799f8746b43001e0f7411eb5d359e17cabc6bfce4a7bccbfdd99",
24
+ "bytes": 844
25
+ },
26
+ "coreai/assets/g2p/normalization.json": {
27
+ "sha256": "f9ca4fe4e4eca5e02b6f2874f5da6ede8b44b25c71c3c6ac00f96d7397024de6",
28
+ "bytes": 1090
29
+ },
30
+ "coreai/assets/g2p/phoneme_dict.json": {
31
+ "sha256": "9cfb40b2fd31d7a32351c33f90230062554a4dc0a44e89f722de211d42ed1875",
32
+ "bytes": 93
33
+ },
34
+ "coreai/assets/g2p/pronunciation_overrides.json": {
35
+ "sha256": "a1f311379b76cb44719a2242a93a4046ff5f344cdac613374288234c2817eb59",
36
+ "bytes": 80
37
+ },
38
+ "coreai/assets/g2p/tokenizer.json": {
39
+ "sha256": "4843d57a422b0a6045bbc6fc9a99afa329126a6742ff95ebcefcda660f3ee474",
40
+ "bytes": 7769
41
+ },
42
+ "coreai/assets/languages.json": {
43
+ "sha256": "f2af2233d0a1c0dbb6e6236fd57592d12da84262fdd1a336ead1ee83a657e2ae",
44
+ "bytes": 289
45
+ },
46
+ "coreai/assets/phone_vocab.json": {
47
+ "sha256": "83e6457a69948edbd9a94b404c22fd408c263eb0d0b5acaafa4926e994587721",
48
+ "bytes": 1793
49
+ },
50
+ "coreai/assets/voices.json": {
51
+ "sha256": "64b1b317b859e8b98d2802ecc2ff5be7edfcb82d4c641a4aa4a5a4d18fc4841b",
52
+ "bytes": 613
53
+ },
54
+ "coreai/coreai/duration_predictor.aimodel/main.hash": {
55
+ "sha256": "2a0d251ee3172c4bcc7453c1ea2f30284f0d7750e9d6b16e19863c15f2db828d",
56
+ "bytes": 32
57
+ },
58
+ "coreai/coreai/duration_predictor.aimodel/main.mlirb": {
59
+ "sha256": "fa30904d7f7cca5cabb4d5b32130522c7a7495ec43f89210154fca66b3ea3f66",
60
+ "bytes": 10921604
61
+ },
62
+ "coreai/coreai/duration_predictor.aimodel/metadata.json": {
63
+ "sha256": "983438b7ab32758ba2538b45a2f4cfba3cce6a9cb7bef24408c8a793ba9b9b96",
64
+ "bytes": 105
65
+ },
66
+ "coreai/coreai/g2p.aimodel/main.hash": {
67
+ "sha256": "c591fb2ea25417c7e1ba2f062425662bdbd781dee25e2b4021c9fe11bab23bdb",
68
+ "bytes": 32
69
+ },
70
+ "coreai/coreai/g2p.aimodel/main.mlirb": {
71
+ "sha256": "612cb67bfc472eb22bceec7796ad3e97e54fb09aedcc545dd91f717e0f36b421",
72
+ "bytes": 28377289
73
+ },
74
+ "coreai/coreai/g2p.aimodel/metadata.json": {
75
+ "sha256": "3b38b50e50ff11f0cb377d3d9d14a86c3d857208be95e12191e501023046fd0d",
76
+ "bytes": 105
77
+ },
78
+ "coreai/coreai/pitch_context_encoder.aimodel/main.hash": {
79
+ "sha256": "f290d3f805bc41a40ef4fd59b7fa045e6cc29a0e0312e158c547950d5029407b",
80
+ "bytes": 32
81
+ },
82
+ "coreai/coreai/pitch_context_encoder.aimodel/main.mlirb": {
83
+ "sha256": "bba55947ab3242f7b80146b3b6247478f6da217d3360b535dc0dfefaec5bd1a9",
84
+ "bytes": 4700833
85
+ },
86
+ "coreai/coreai/pitch_context_encoder.aimodel/metadata.json": {
87
+ "sha256": "b7725bd090086f801f99bab0b7101400252c9d1bed22d8ba9fd07082250dcaf0",
88
+ "bytes": 105
89
+ },
90
+ "coreai/coreai/pitch_estimator.aimodel/main.hash": {
91
+ "sha256": "09709b7a70e46913f59d9c6476cab1ce55a6ad22844b2db6d4a1ef05df5659fa",
92
+ "bytes": 32
93
+ },
94
+ "coreai/coreai/pitch_estimator.aimodel/main.mlirb": {
95
+ "sha256": "546c6f84e167b61b37553686e6f149aa6d49a60b888d99cd058aecc8619be142",
96
+ "bytes": 25681787
97
+ },
98
+ "coreai/coreai/pitch_estimator.aimodel/metadata.json": {
99
+ "sha256": "f116d2ca426a5f27c223ca92b8f8a2885c67cc8d75ca7ab3dbb05ce1a7f2c501",
100
+ "bytes": 105
101
+ },
102
+ "coreai/coreai/vector_context_encoder.aimodel/main.hash": {
103
+ "sha256": "2b8f0c248fea221d5ff3149b6e17eca1499763ddc90beb10997bcf4c3b505ee2",
104
+ "bytes": 32
105
+ },
106
+ "coreai/coreai/vector_context_encoder.aimodel/main.mlirb": {
107
+ "sha256": "137937a27b388dd8788234a5681888ed81168e12de7be346c5d0de0c13db2b8e",
108
+ "bytes": 8029278
109
+ },
110
+ "coreai/coreai/vector_context_encoder.aimodel/metadata.json": {
111
+ "sha256": "8f28f1781c428e199c0799793d77462ca30f83dadc03058d5cef2351efe1d01c",
112
+ "bytes": 105
113
+ },
114
+ "coreai/coreai/vector_estimator.aimodel/main.hash": {
115
+ "sha256": "37adbc7923acc6c69868fedd504397984907c3c2d053f97ca76f82e84e7dd7d2",
116
+ "bytes": 32
117
+ },
118
+ "coreai/coreai/vector_estimator.aimodel/main.mlirb": {
119
+ "sha256": "664969ffee72ee62dc91897ddb0a237c984e6156b8ed9d5a6356ae868914894d",
120
+ "bytes": 63194410
121
+ },
122
+ "coreai/coreai/vector_estimator.aimodel/metadata.json": {
123
+ "sha256": "c9012dac492169425ad672734e09f841e55dd0b5d8e7bca9bd2b71f3f397bce4",
124
+ "bytes": 105
125
+ },
126
+ "coreai/coreai/vocoder.aimodel/main.hash": {
127
+ "sha256": "ce363368896e6e396519a60d7003ac2ee3cd47101865642015e409a6a2659056",
128
+ "bytes": 32
129
+ },
130
+ "coreai/coreai/vocoder.aimodel/main.mlirb": {
131
+ "sha256": "62bd2889536a7c3de0d7856259c0a2f5d2cb7f6d8834fce0fb6994aea87c02da",
132
+ "bytes": 43814084
133
+ },
134
+ "coreai/coreai/vocoder.aimodel/metadata.json": {
135
+ "sha256": "305c972524d434d7e1cfb795ba5e10137dc853056ae332cfd0e79da49511cd82",
136
+ "bytes": 105
137
+ },
138
+ "coreai/manifest.json": {
139
+ "sha256": "05fd3f07a11c40a4f29d2f27d8753911988f2000e47fd218aa9bb329b4784bc7",
140
+ "bytes": 32451
141
+ },
142
+ "coreml/assets/fixed_references.json": {
143
+ "sha256": "f227c55cab5f6b8cffa33c3dd743e69b26c7e446be5b9cd9a286095cabaa249e",
144
+ "bytes": 1555673
145
+ },
146
+ "coreml/assets/g2p/export_config.json": {
147
+ "sha256": "0ec5ea81f346521f3a5c9c4a90b939c6b5143b1a11cc5e74b9902969db56e568",
148
+ "bytes": 928
149
+ },
150
+ "coreml/assets/g2p/language_map.json": {
151
+ "sha256": "b3db2048b492799f8746b43001e0f7411eb5d359e17cabc6bfce4a7bccbfdd99",
152
+ "bytes": 844
153
+ },
154
+ "coreml/assets/g2p/normalization.json": {
155
+ "sha256": "f9ca4fe4e4eca5e02b6f2874f5da6ede8b44b25c71c3c6ac00f96d7397024de6",
156
+ "bytes": 1090
157
+ },
158
+ "coreml/assets/g2p/phoneme_dict.json": {
159
+ "sha256": "9cfb40b2fd31d7a32351c33f90230062554a4dc0a44e89f722de211d42ed1875",
160
+ "bytes": 93
161
+ },
162
+ "coreml/assets/g2p/pronunciation_overrides.json": {
163
+ "sha256": "a1f311379b76cb44719a2242a93a4046ff5f344cdac613374288234c2817eb59",
164
+ "bytes": 80
165
+ },
166
+ "coreml/assets/g2p/tokenizer.json": {
167
+ "sha256": "4843d57a422b0a6045bbc6fc9a99afa329126a6742ff95ebcefcda660f3ee474",
168
+ "bytes": 7769
169
+ },
170
+ "coreml/assets/languages.json": {
171
+ "sha256": "f2af2233d0a1c0dbb6e6236fd57592d12da84262fdd1a336ead1ee83a657e2ae",
172
+ "bytes": 289
173
+ },
174
+ "coreml/assets/phone_vocab.json": {
175
+ "sha256": "83e6457a69948edbd9a94b404c22fd408c263eb0d0b5acaafa4926e994587721",
176
+ "bytes": 1793
177
+ },
178
+ "coreml/assets/voices.json": {
179
+ "sha256": "64b1b317b859e8b98d2802ecc2ff5be7edfcb82d4c641a4aa4a5a4d18fc4841b",
180
+ "bytes": 613
181
+ },
182
+ "coreml/coreml/duration_predictor.mlmodelc/analytics/coremldata.bin": {
183
+ "sha256": "ff481a2cd37fa03e16ecda507b033ec1f76b69469819a3259e3574bf9ebe72ab",
184
+ "bytes": 243
185
+ },
186
+ "coreml/coreml/duration_predictor.mlmodelc/coremldata.bin": {
187
+ "sha256": "06bfdf64690cc0de4e97a0af486c678bf498b4b4224159fd40af1ba8ae296c60",
188
+ "bytes": 554
189
+ },
190
+ "coreml/coreml/duration_predictor.mlmodelc/model.mil": {
191
+ "sha256": "3287da9167c504b94822662cbbf78d95815ae2a173bc34f2265b9079bb1936e0",
192
+ "bytes": 139155
193
+ },
194
+ "coreml/coreml/duration_predictor.mlmodelc/weights/weight.bin": {
195
+ "sha256": "6e6314ef13cc2edaffdc96840a5bf4c7e729630fb22e562f5383be48e3458038",
196
+ "bytes": 10580928
197
+ },
198
+ "coreml/coreml/g2p.mlmodelc/analytics/coremldata.bin": {
199
+ "sha256": "9047c70b85b77b6d386f1e6188323f82daf023854c2ae4c7018f8406c04320fb",
200
+ "bytes": 243
201
+ },
202
+ "coreml/coreml/g2p.mlmodelc/coremldata.bin": {
203
+ "sha256": "f5fc8bee0186f33fa14d35a59a58ec3393676d07bdf03802a3836d806662cd03",
204
+ "bytes": 377
205
+ },
206
+ "coreml/coreml/g2p.mlmodelc/model.mil": {
207
+ "sha256": "d464334a231d99b9a3764d50c28657d1720e25eaa45e39fa4daa401bb89452a1",
208
+ "bytes": 388409
209
+ },
210
+ "coreml/coreml/g2p.mlmodelc/weights/weight.bin": {
211
+ "sha256": "045da2991ed2a42ba027aaa6b83813054f410b6eedfdce82f2ab287ec4ca74e2",
212
+ "bytes": 27346752
213
+ },
214
+ "coreml/coreml/pitch_context_encoder.mlmodelc/analytics/coremldata.bin": {
215
+ "sha256": "7ef86aaa3ec78ce0ff02c2927c02dcb1d073d64ddd6b78e4358af91cbd0842d2",
216
+ "bytes": 243
217
+ },
218
+ "coreml/coreml/pitch_context_encoder.mlmodelc/coremldata.bin": {
219
+ "sha256": "9e77e8333bd4ccb529ba0b96fc342cece9114686fc381837dac32e354a5ee50d",
220
+ "bytes": 476
221
+ },
222
+ "coreml/coreml/pitch_context_encoder.mlmodelc/model.mil": {
223
+ "sha256": "219c734dc5969aa5355bfff1af22eb8b488907a2e84c36407a13c6a30e9304a9",
224
+ "bytes": 44801
225
+ },
226
+ "coreml/coreml/pitch_context_encoder.mlmodelc/weights/weight.bin": {
227
+ "sha256": "240f333dbe5c13044acae1642fb7985ea0db3f0495874851bd78228f4c3cd334",
228
+ "bytes": 4353280
229
+ },
230
+ "coreml/coreml/pitch_estimator.mlmodelc/analytics/coremldata.bin": {
231
+ "sha256": "d4fa8be0476ebae87fe524d6e640a2563e6807d226f14ca8b3f63faebfe566f5",
232
+ "bytes": 243
233
+ },
234
+ "coreml/coreml/pitch_estimator.mlmodelc/coremldata.bin": {
235
+ "sha256": "afeb64f3c064a15518f25027ab0ebd28b09f8768ac8a0f70df65f542f0bd8bdf",
236
+ "bytes": 3292
237
+ },
238
+ "coreml/coreml/pitch_estimator.mlmodelc/model.mil": {
239
+ "sha256": "2973fd3c7a6f93d0166c2adefc5e3fa5984f0161cad025e349d2243482b2380e",
240
+ "bytes": 1039480
241
+ },
242
+ "coreml/coreml/pitch_estimator.mlmodelc/weights/weight.bin": {
243
+ "sha256": "1df3ce395d51695966dfca278714e6aa9a5641a4e2c2c5c8c2a3defbbf12683f",
244
+ "bytes": 23927936
245
+ },
246
+ "coreml/coreml/vector_context_encoder.mlmodelc/analytics/coremldata.bin": {
247
+ "sha256": "1f5a434cb6dc9d6137bc8a5c26bf7b49716c35a4e6b252b104f03a0fb5b53c2d",
248
+ "bytes": 243
249
+ },
250
+ "coreml/coreml/vector_context_encoder.mlmodelc/coremldata.bin": {
251
+ "sha256": "2831a5db7df542f88d581caec0b3a1091cd6e2522c0a7c82da21357327719517",
252
+ "bytes": 476
253
+ },
254
+ "coreml/coreml/vector_context_encoder.mlmodelc/model.mil": {
255
+ "sha256": "c3c0df163512f0ed8ab4dbb927054f2cb627980c734419bd96e341540660d372",
256
+ "bytes": 44809
257
+ },
258
+ "coreml/coreml/vector_context_encoder.mlmodelc/weights/weight.bin": {
259
+ "sha256": "f686199febbe315c2102abd521d33fb629813d2293afe9daa08c94828d14d66e",
260
+ "bytes": 7572352
261
+ },
262
+ "coreml/coreml/vector_estimator.mlmodelc/analytics/coremldata.bin": {
263
+ "sha256": "4a60369557fe37fcd46f6245d6de4d4a55806967f5dfbca510a75e968a826882",
264
+ "bytes": 243
265
+ },
266
+ "coreml/coreml/vector_estimator.mlmodelc/coremldata.bin": {
267
+ "sha256": "e9b0d967ca6ab8cace2814a61e7c1561d9a45c7b65b2373825b6d53e10b9b47f",
268
+ "bytes": 3845
269
+ },
270
+ "coreml/coreml/vector_estimator.mlmodelc/model.mil": {
271
+ "sha256": "54958da9f797a1ca4f92d633e31e153c0330e445804797db2913653816b830a3",
272
+ "bytes": 1474942
273
+ },
274
+ "coreml/coreml/vector_estimator.mlmodelc/weights/weight.bin": {
275
+ "sha256": "b808c10f9838b2bbf671ccef3d765cebabf58243457a01c06ec11480a9392388",
276
+ "bytes": 60779072
277
+ },
278
+ "coreml/coreml/vocoder.mlmodelc/analytics/coremldata.bin": {
279
+ "sha256": "60f03103f36e751ddfa285184d62a93c1edc615d5fa9833e48ccc93ed577464d",
280
+ "bytes": 243
281
+ },
282
+ "coreml/coreml/vocoder.mlmodelc/coremldata.bin": {
283
+ "sha256": "04bf9f9de7edfc6ce58e5928a0a1d6669d09095ebc03cb5ca24dd85a416653df",
284
+ "bytes": 1190
285
+ },
286
+ "coreml/coreml/vocoder.mlmodelc/model.mil": {
287
+ "sha256": "8e5c9ef45e5d10703d6af341438c78789b0fd17659b6a074194aac05eb826d6c",
288
+ "bytes": 970273
289
+ },
290
+ "coreml/coreml/vocoder.mlmodelc/weights/weight.bin": {
291
+ "sha256": "4f787f1a85ba0b25deb1baa8144e3bb8cf4095a5c16690e2e0700cc53402dcf5",
292
+ "bytes": 31041408
293
+ },
294
+ "coreml/manifest.json": {
295
+ "sha256": "7e227f026716ad7dcf24f013f71dcb2504d05d6f09850700afef0d8a8e364b41",
296
+ "bytes": 32276
297
+ },
298
  "litert/assets/fixed_references.json": {
299
  "sha256": "f227c55cab5f6b8cffa33c3dd743e69b26c7e446be5b9cd9a286095cabaa249e",
300
  "bytes": 1555673
coreai/assets/fixed_references.json ADDED
The diff for this file is too large to render. See raw diff
 
coreai/assets/g2p/export_config.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "batch_dynamic": true,
3
+ "format": "scyllasband_g2p",
4
+ "input_granularity": "phrase",
5
+ "languages": [
6
+ "de",
7
+ "en_gb",
8
+ "en_us",
9
+ "es",
10
+ "fr",
11
+ "it",
12
+ "vi"
13
+ ],
14
+ "model_type": "transformer",
15
+ "onnx_export_batch_size": 4,
16
+ "onnx_export_layout": "batch_first",
17
+ "onnx_export_max_batch_size": 256,
18
+ "onnx_exporter": "dynamo",
19
+ "onnx_fp32_model": "model.onnx",
20
+ "onnx_model": "model.onnx",
21
+ "onnx_padded_output": "ctc_blank",
22
+ "onnx_quantization": "none",
23
+ "onnx_quantized_parity_tolerance": 3.0,
24
+ "opset": 18,
25
+ "phoneme_token_mode": "char",
26
+ "phrase_rows_per_language": 200000,
27
+ "preserve_commas": true,
28
+ "punctuation_silence_target": "explicit_silence",
29
+ "release_role": "fp32_bundle_frontend",
30
+ "source_checkpoint": "output/deepphonemizer_multilingual/model_v10_7lang/checkpoints/model_step_250k.pt",
31
+ "training_mix": "phrase_word_chunked_clean",
32
+ "word_rows_per_language": 200000
33
+ }
coreai/assets/g2p/language_map.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "american": "en_us",
3
+ "american_english": "en_us",
4
+ "british": "en_gb",
5
+ "british_english": "en_gb",
6
+ "de": "de",
7
+ "de_de": "de",
8
+ "deu": "de",
9
+ "en": "en_us",
10
+ "en-gb": "en_gb",
11
+ "en-uk": "en_gb",
12
+ "en-us": "en_us",
13
+ "en_gb": "en_gb",
14
+ "en_uk": "en_gb",
15
+ "en_us": "en_us",
16
+ "eng": "en_us",
17
+ "eng_gb": "en_gb",
18
+ "eng_us": "en_us",
19
+ "english": "en_us",
20
+ "english_gb": "en_gb",
21
+ "english_uk": "en_gb",
22
+ "english_us": "en_us",
23
+ "es": "es",
24
+ "es_es": "es",
25
+ "es_mx": "es",
26
+ "fr": "fr",
27
+ "fr_fr": "fr",
28
+ "fra": "fr",
29
+ "fre": "fr",
30
+ "french": "fr",
31
+ "gb": "en_gb",
32
+ "ger": "de",
33
+ "german": "de",
34
+ "it": "it",
35
+ "it_it": "it",
36
+ "ita": "it",
37
+ "italian": "it",
38
+ "spa": "es",
39
+ "spanish": "es",
40
+ "uk": "en_gb",
41
+ "us": "en_us",
42
+ "vi": "vi",
43
+ "vi_hn": "vi",
44
+ "vi_vn": "vi",
45
+ "vie": "vi",
46
+ "vietnamese": "vi"
47
+ }
coreai/assets/g2p/normalization.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "contract": {
3
+ "colon_semicolon_classes": "preserved",
4
+ "dot_runs_two_or_more": "ellipsis_three_dots",
5
+ "homogeneous_terminal_runs": "single_mark",
6
+ "lexical_hyphen": "space_no_pause",
7
+ "mixed_question_exclamation": "leading_mark",
8
+ "non_acoustic_modifiers": [
9
+ "stress",
10
+ "length",
11
+ "unicode_combining"
12
+ ],
13
+ "ok_spoken_form": "okay",
14
+ "punctuation_silence_target": "explicit_silence",
15
+ "syntactic_dash": "em_dash_pause",
16
+ "training_phone_source": "espeak",
17
+ "version": "scyllasband_spoken_text_v2_2026_08_05"
18
+ },
19
+ "contract_sha256": "d8fd1b2415d0d100d1c70237a4469117ee6b73d0e4da703e4654c83192779cc4",
20
+ "contract_version": "scyllasband_spoken_text_v2_2026_08_05",
21
+ "expand_currency": true,
22
+ "expand_dates": true,
23
+ "expand_numbers": true,
24
+ "expand_ordinals": true,
25
+ "expand_percentages": true,
26
+ "expand_times": true,
27
+ "languages": [
28
+ "en",
29
+ "en_us",
30
+ "en_gb",
31
+ "es",
32
+ "it",
33
+ "fr",
34
+ "de",
35
+ "vi"
36
+ ],
37
+ "normalize_at_sign": true,
38
+ "normalize_punctuation": true,
39
+ "normalizer": "SpokenTextNormalizer"
40
+ }
coreai/assets/g2p/phoneme_dict.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "de": {},
3
+ "en_gb": {},
4
+ "en_us": {},
5
+ "es": {},
6
+ "fr": {},
7
+ "it": {},
8
+ "vi": {}
9
+ }
coreai/assets/g2p/pronunciation_overrides.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "format": "scyllasband_g2p_pronunciation_overrides_v1",
3
+ "languages": {}
4
+ }
coreai/assets/g2p/tokenizer.json ADDED
@@ -0,0 +1,517 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "format": "deepphonemizer_forward_ctc",
3
+ "languages": [
4
+ "de",
5
+ "en_gb",
6
+ "en_us",
7
+ "es",
8
+ "fr",
9
+ "it",
10
+ "vi"
11
+ ],
12
+ "char_repeats": 4,
13
+ "lowercase": true,
14
+ "phoneme_token_mode": "char",
15
+ "text_symbols": {
16
+ "_": 0,
17
+ "<de>": 1,
18
+ "<en_gb>": 2,
19
+ "<en_us>": 3,
20
+ "<es>": 4,
21
+ "<fr>": 5,
22
+ "<it>": 6,
23
+ "<vi>": 7,
24
+ "<end>": 8,
25
+ "\b": 9,
26
+ " ": 10,
27
+ "!": 11,
28
+ "\"": 12,
29
+ "#": 13,
30
+ "$": 14,
31
+ "%": 15,
32
+ "&": 16,
33
+ "'": 17,
34
+ "(": 18,
35
+ ")": 19,
36
+ "*": 20,
37
+ "+": 21,
38
+ ",": 22,
39
+ "-": 23,
40
+ ".": 24,
41
+ "/": 25,
42
+ "0": 26,
43
+ "1": 27,
44
+ "2": 28,
45
+ "3": 29,
46
+ "4": 30,
47
+ "5": 31,
48
+ "6": 32,
49
+ "7": 33,
50
+ "8": 34,
51
+ "9": 35,
52
+ ":": 36,
53
+ ";": 37,
54
+ "<": 38,
55
+ "=": 39,
56
+ ">": 40,
57
+ "?": 41,
58
+ "@": 42,
59
+ "A": 43,
60
+ "B": 44,
61
+ "C": 45,
62
+ "D": 46,
63
+ "E": 47,
64
+ "F": 48,
65
+ "G": 49,
66
+ "H": 50,
67
+ "I": 51,
68
+ "J": 52,
69
+ "K": 53,
70
+ "L": 54,
71
+ "M": 55,
72
+ "N": 56,
73
+ "O": 57,
74
+ "P": 58,
75
+ "Q": 59,
76
+ "R": 60,
77
+ "S": 61,
78
+ "T": 62,
79
+ "U": 63,
80
+ "V": 64,
81
+ "W": 65,
82
+ "X": 66,
83
+ "Y": 67,
84
+ "Z": 68,
85
+ "[": 69,
86
+ "\\": 70,
87
+ "]": 71,
88
+ "^": 72,
89
+ "`": 73,
90
+ "a": 74,
91
+ "b": 75,
92
+ "c": 76,
93
+ "d": 77,
94
+ "e": 78,
95
+ "f": 79,
96
+ "g": 80,
97
+ "h": 81,
98
+ "i": 82,
99
+ "j": 83,
100
+ "k": 84,
101
+ "l": 85,
102
+ "m": 86,
103
+ "n": 87,
104
+ "o": 88,
105
+ "p": 89,
106
+ "q": 90,
107
+ "r": 91,
108
+ "s": 92,
109
+ "t": 93,
110
+ "u": 94,
111
+ "v": 95,
112
+ "w": 96,
113
+ "x": 97,
114
+ "y": 98,
115
+ "z": 99,
116
+ "{": 100,
117
+ "|": 101,
118
+ "}": 102,
119
+ "~": 103,
120
+ "¡": 104,
121
+ "¢": 105,
122
+ "£": 106,
123
+ "¥": 107,
124
+ "§": 108,
125
+ "ª": 109,
126
+ "«": 110,
127
+ "­": 111,
128
+ "®": 112,
129
+ "°": 113,
130
+ "±": 114,
131
+ "²": 115,
132
+ "³": 116,
133
+ "´": 117,
134
+ "¶": 118,
135
+ "·": 119,
136
+ "¹": 120,
137
+ "º": 121,
138
+ "»": 122,
139
+ "¼": 123,
140
+ "½": 124,
141
+ "¾": 125,
142
+ "¿": 126,
143
+ "À": 127,
144
+ "Á": 128,
145
+ "Â": 129,
146
+ "Ä": 130,
147
+ "Å": 131,
148
+ "Æ": 132,
149
+ "Ç": 133,
150
+ "È": 134,
151
+ "É": 135,
152
+ "Ê": 136,
153
+ "Í": 137,
154
+ "Ï": 138,
155
+ "Ñ": 139,
156
+ "Ó": 140,
157
+ "Ô": 141,
158
+ "Ö": 142,
159
+ "×": 143,
160
+ "Ù": 144,
161
+ "Ú": 145,
162
+ "Ü": 146,
163
+ "Ý": 147,
164
+ "Þ": 148,
165
+ "ß": 149,
166
+ "à": 150,
167
+ "á": 151,
168
+ "â": 152,
169
+ "ã": 153,
170
+ "ä": 154,
171
+ "å": 155,
172
+ "æ": 156,
173
+ "ç": 157,
174
+ "è": 158,
175
+ "é": 159,
176
+ "ê": 160,
177
+ "ë": 161,
178
+ "ì": 162,
179
+ "í": 163,
180
+ "î": 164,
181
+ "ï": 165,
182
+ "ð": 166,
183
+ "ñ": 167,
184
+ "ò": 168,
185
+ "ó": 169,
186
+ "ô": 170,
187
+ "õ": 171,
188
+ "ö": 172,
189
+ "ø": 173,
190
+ "ù": 174,
191
+ "ú": 175,
192
+ "û": 176,
193
+ "ü": 177,
194
+ "ý": 178,
195
+ "þ": 179,
196
+ "ÿ": 180,
197
+ "ā": 181,
198
+ "Ă": 182,
199
+ "ă": 183,
200
+ "ą": 184,
201
+ "ć": 185,
202
+ "č": 186,
203
+ "Đ": 187,
204
+ "đ": 188,
205
+ "ē": 189,
206
+ "ė": 190,
207
+ "ę": 191,
208
+ "ğ": 192,
209
+ "ġ": 193,
210
+ "ĩ": 194,
211
+ "ī": 195,
212
+ "ļ": 196,
213
+ "ł": 197,
214
+ "ń": 198,
215
+ "Ō": 199,
216
+ "ō": 200,
217
+ "Œ": 201,
218
+ "œ": 202,
219
+ "ŕ": 203,
220
+ "ś": 204,
221
+ "ş": 205,
222
+ "š": 206,
223
+ "ŧ": 207,
224
+ "ũ": 208,
225
+ "ū": 209,
226
+ "ŭ": 210,
227
+ "ů": 211,
228
+ "ű": 212,
229
+ "ż": 213,
230
+ "Ž": 214,
231
+ "ž": 215,
232
+ "ſ": 216,
233
+ "Ơ": 217,
234
+ "ơ": 218,
235
+ "Ư": 219,
236
+ "ư": 220,
237
+ "ǁ": 221,
238
+ "ǂ": 222,
239
+ "ǐ": 223,
240
+ "ǘ": 224,
241
+ "ǧ": 225,
242
+ "ș": 226,
243
+ "Ȝ": 227,
244
+ "ȝ": 228,
245
+ "ȟ": 229,
246
+ "ɛ": 230,
247
+ "ʰ": 231,
248
+ "ˆ": 232,
249
+ "ˈ": 233,
250
+ "ˢ": 234,
251
+ "̃": 235,
252
+ "̍": 236,
253
+ "̣": 237,
254
+ "̱": 238,
255
+ "͈": 239,
256
+ "α": 240,
257
+ "β": 241,
258
+ "γ": 242,
259
+ "δ": 243,
260
+ "ε": 244,
261
+ "η": 245,
262
+ "κ": 246,
263
+ "μ": 247,
264
+ "ξ": 248,
265
+ "π": 249,
266
+ "ς": 250,
267
+ "σ": 251,
268
+ "χ": 252,
269
+ "ψ": 253,
270
+ "а": 254,
271
+ "д": 255,
272
+ "е": 256,
273
+ "з": 257,
274
+ "и": 258,
275
+ "к": 259,
276
+ "л": 260,
277
+ "о": 261,
278
+ "п": 262,
279
+ "р": 263,
280
+ "с": 264,
281
+ "т": 265,
282
+ "ч": 266,
283
+ "ь": 267,
284
+ "ง": 268,
285
+ "ย": 269,
286
+ "า": 270,
287
+ "่": 271,
288
+ "ည": 272,
289
+ "လ": 273,
290
+ "း": 274,
291
+ "်": 275,
292
+ "ᵈ": 276,
293
+ "ᵗ": 277,
294
+ "ḅ": 278,
295
+ "ḍ": 279,
296
+ "ḳ": 280,
297
+ "ṅ": 281,
298
+ "Ṭ": 282,
299
+ "ẓ": 283,
300
+ "ạ": 284,
301
+ "Ả": 285,
302
+ "ả": 286,
303
+ "Ấ": 287,
304
+ "ấ": 288,
305
+ "ầ": 289,
306
+ "ẩ": 290,
307
+ "ẫ": 291,
308
+ "ậ": 292,
309
+ "ắ": 293,
310
+ "ằ": 294,
311
+ "ẳ": 295,
312
+ "ẵ": 296,
313
+ "ặ": 297,
314
+ "ẹ": 298,
315
+ "ẻ": 299,
316
+ "ẽ": 300,
317
+ "Ế": 301,
318
+ "ế": 302,
319
+ "Ề": 303,
320
+ "ề": 304,
321
+ "ể": 305,
322
+ "ễ": 306,
323
+ "ệ": 307,
324
+ "ỉ": 308,
325
+ "ị": 309,
326
+ "ọ": 310,
327
+ "ỏ": 311,
328
+ "Ố": 312,
329
+ "ố": 313,
330
+ "Ồ": 314,
331
+ "ồ": 315,
332
+ "Ổ": 316,
333
+ "ổ": 317,
334
+ "ỗ": 318,
335
+ "Ộ": 319,
336
+ "ộ": 320,
337
+ "ớ": 321,
338
+ "Ờ": 322,
339
+ "ờ": 323,
340
+ "Ở": 324,
341
+ "ở": 325,
342
+ "ỡ": 326,
343
+ "ợ": 327,
344
+ "ụ": 328,
345
+ "ủ": 329,
346
+ "Ứ": 330,
347
+ "ứ": 331,
348
+ "Ừ": 332,
349
+ "ừ": 333,
350
+ "ử": 334,
351
+ "ữ": 335,
352
+ "ự": 336,
353
+ "ỳ": 337,
354
+ "ỵ": 338,
355
+ "Ỷ": 339,
356
+ "ỷ": 340,
357
+ "ỹ": 341,
358
+ "ὖ": 342,
359
+ "ὼ": 343,
360
+ "ῆ": 344,
361
+ "​": 345,
362
+ "‐": 346,
363
+ "—": 347,
364
+ "―": 348,
365
+ "†": 349,
366
+ "•": 350,
367
+ "…": 351,
368
+ "‰": 352,
369
+ "′": 353,
370
+ "″": 354,
371
+ "⁄": 355,
372
+ "⁷": 356,
373
+ "⁻": 357,
374
+ "₀": 358,
375
+ "₁": 359,
376
+ "₂": 360,
377
+ "₃": 361,
378
+ "₄": 362,
379
+ "₅": 363,
380
+ "₆": 364,
381
+ "₇": 365,
382
+ "₈": 366,
383
+ "€": 367,
384
+ "₹": 368,
385
+ "⅓": 369,
386
+ "⅔": 370,
387
+ "⅖": 371,
388
+ "⅙": 372,
389
+ "⅛": 373,
390
+ "⅝": 374,
391
+ "√": 375,
392
+ "│": 376,
393
+ "♦": 377,
394
+ "⚓": 378,
395
+ "✈": 379,
396
+ "✠": 380,
397
+ "✨": 381,
398
+ "⟙": 382,
399
+ "⸺": 383,
400
+ "。": 384,
401
+ "〈": 385,
402
+ "な": 386,
403
+ "三": 387,
404
+ "次": 388,
405
+ "肿": 389,
406
+ "臃": 390,
407
+ "구": 391,
408
+ "로": 392,
409
+ "요": 393,
410
+ "ﷺ": 394,
411
+ "️": 395,
412
+ "𝑥": 396,
413
+ "𝑦": 397,
414
+ "🌊": 398,
415
+ "🌿": 399,
416
+ "💦": 400,
417
+ "💨": 401
418
+ },
419
+ "phoneme_symbols": {
420
+ "0": "_",
421
+ "1": "<de>",
422
+ "2": "<en_gb>",
423
+ "3": "<en_us>",
424
+ "4": "<es>",
425
+ "5": "<fr>",
426
+ "6": "<it>",
427
+ "7": "<vi>",
428
+ "8": "<end>",
429
+ "9": " ",
430
+ "10": "!",
431
+ "11": ",",
432
+ "12": "-",
433
+ "13": ".",
434
+ "14": "1",
435
+ "15": "2",
436
+ "16": "4",
437
+ "17": "5",
438
+ "18": "6",
439
+ "19": "7",
440
+ "20": ":",
441
+ "21": ";",
442
+ "22": "?",
443
+ "23": "a",
444
+ "24": "b",
445
+ "25": "c",
446
+ "26": "d",
447
+ "27": "e",
448
+ "28": "f",
449
+ "29": "h",
450
+ "30": "i",
451
+ "31": "j",
452
+ "32": "k",
453
+ "33": "l",
454
+ "34": "m",
455
+ "35": "n",
456
+ "36": "o",
457
+ "37": "p",
458
+ "38": "q",
459
+ "39": "r",
460
+ "40": "s",
461
+ "41": "t",
462
+ "42": "u",
463
+ "43": "v",
464
+ "44": "w",
465
+ "45": "x",
466
+ "46": "y",
467
+ "47": "z",
468
+ "48": "«",
469
+ "49": "»",
470
+ "50": "æ",
471
+ "51": "ç",
472
+ "52": "ð",
473
+ "53": "ø",
474
+ "54": "ŋ",
475
+ "55": "œ",
476
+ "56": "ɐ",
477
+ "57": "ɑ",
478
+ "58": "ɒ",
479
+ "59": "ɔ",
480
+ "60": "ɗ",
481
+ "61": "ə",
482
+ "62": "ɚ",
483
+ "63": "ɛ",
484
+ "64": "ɜ",
485
+ "65": "ɟ",
486
+ "66": "ɡ",
487
+ "67": "ɣ",
488
+ "68": "ɨ",
489
+ "69": "ɪ",
490
+ "70": "ɬ",
491
+ "71": "ɯ",
492
+ "72": "ɲ",
493
+ "73": "ɹ",
494
+ "74": "ɾ",
495
+ "75": "ʁ",
496
+ "76": "ʃ",
497
+ "77": "ʊ",
498
+ "78": "ʌ",
499
+ "79": "ʎ",
500
+ "80": "ʐ",
501
+ "81": "ʒ",
502
+ "82": "ʔ",
503
+ "83": "ʝ",
504
+ "84": "ʲ",
505
+ "85": "ˈ",
506
+ "86": "ˌ",
507
+ "87": "ː",
508
+ "88": "̃",
509
+ "89": "̩",
510
+ "90": "̪",
511
+ "91": "β",
512
+ "92": "θ",
513
+ "93": "ᵻ",
514
+ "94": "—",
515
+ "95": "…"
516
+ }
517
+ }
coreai/assets/languages.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "id": "en_us",
4
+ "index": 0
5
+ },
6
+ {
7
+ "id": "en_gb",
8
+ "index": 1
9
+ },
10
+ {
11
+ "id": "es",
12
+ "index": 2
13
+ },
14
+ {
15
+ "id": "it",
16
+ "index": 3
17
+ },
18
+ {
19
+ "id": "fr",
20
+ "index": 4
21
+ },
22
+ {
23
+ "id": "de",
24
+ "index": 5
25
+ },
26
+ {
27
+ "id": "vi",
28
+ "index": 6
29
+ }
30
+ ]
coreai/assets/phone_vocab.json ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "mode": "scyllasband_ipa_phone_symbols",
4
+ "special_tokens": [
5
+ "<pad>",
6
+ "<unk>",
7
+ "<sil>",
8
+ "<pause_comma>",
9
+ "<pause_semicolon>",
10
+ "<pause_colon>",
11
+ "<pause_dash>",
12
+ "<ellipsis>",
13
+ "<end_stmt>",
14
+ "<end_question>",
15
+ "<end_exclaim>",
16
+ "<ctx_sentence_start>",
17
+ "<ctx_continuation>",
18
+ "<ctx_sentence_end>",
19
+ "<ctx_chunk_continue>"
20
+ ],
21
+ "vocab_size": 87,
22
+ "token_to_id": {
23
+ "<pad>": 0,
24
+ "<unk>": 1,
25
+ "<sil>": 2,
26
+ "<pause_comma>": 3,
27
+ "<pause_semicolon>": 4,
28
+ "<pause_colon>": 5,
29
+ "<pause_dash>": 6,
30
+ "<ellipsis>": 7,
31
+ "<end_stmt>": 8,
32
+ "<end_question>": 9,
33
+ "<end_exclaim>": 10,
34
+ "<ctx_sentence_start>": 11,
35
+ "<ctx_continuation>": 12,
36
+ "<ctx_sentence_end>": 13,
37
+ "<ctx_chunk_continue>": 14,
38
+ "1": 15,
39
+ "2": 16,
40
+ "4": 17,
41
+ "5": 18,
42
+ "6": 19,
43
+ "7": 20,
44
+ "a": 21,
45
+ "b": 22,
46
+ "c": 23,
47
+ "d": 24,
48
+ "e": 25,
49
+ "f": 26,
50
+ "h": 27,
51
+ "i": 28,
52
+ "j": 29,
53
+ "k": 30,
54
+ "l": 31,
55
+ "m": 32,
56
+ "n": 33,
57
+ "o": 34,
58
+ "p": 35,
59
+ "r": 36,
60
+ "s": 37,
61
+ "t": 38,
62
+ "u": 39,
63
+ "v": 40,
64
+ "w": 41,
65
+ "x": 42,
66
+ "y": 43,
67
+ "z": 44,
68
+ "æ": 45,
69
+ "ç": 46,
70
+ "ð": 47,
71
+ "ø": 48,
72
+ "ŋ": 49,
73
+ "œ": 50,
74
+ "ɐ": 51,
75
+ "ɑ": 52,
76
+ "ɒ": 53,
77
+ "ɔ": 54,
78
+ "ɗ": 55,
79
+ "ə": 56,
80
+ "ɚ": 57,
81
+ "ɛ": 58,
82
+ "ɜ": 59,
83
+ "ɟ": 60,
84
+ "ɡ": 61,
85
+ "ɣ": 62,
86
+ "ɪ": 63,
87
+ "ɯ": 64,
88
+ "ɲ": 65,
89
+ "ɹ": 66,
90
+ "ɾ": 67,
91
+ "ʁ": 68,
92
+ "ʃ": 69,
93
+ "ʊ": 70,
94
+ "ʌ": 71,
95
+ "ʎ": 72,
96
+ "ʐ": 73,
97
+ "ʒ": 74,
98
+ "ʔ": 75,
99
+ "ʝ": 76,
100
+ "ʲ": 77,
101
+ "ˈ": 78,
102
+ "ˌ": 79,
103
+ "ː": 80,
104
+ "̃": 81,
105
+ "̩": 82,
106
+ "̪": 83,
107
+ "β": 84,
108
+ "θ": 85,
109
+ "ᵻ": 86
110
+ }
111
+ }
coreai/assets/voices.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "id": "ariadne",
4
+ "index": 0
5
+ },
6
+ {
7
+ "id": "felix",
8
+ "index": 1
9
+ },
10
+ {
11
+ "id": "gwen",
12
+ "index": 2
13
+ },
14
+ {
15
+ "id": "ink",
16
+ "index": 3
17
+ },
18
+ {
19
+ "id": "max",
20
+ "index": 4
21
+ },
22
+ {
23
+ "id": "orpheus",
24
+ "index": 5
25
+ },
26
+ {
27
+ "id": "rex",
28
+ "index": 6
29
+ },
30
+ {
31
+ "id": "scylla",
32
+ "index": 7
33
+ },
34
+ {
35
+ "id": "stone",
36
+ "index": 8
37
+ },
38
+ {
39
+ "id": "tuesday",
40
+ "index": 9
41
+ },
42
+ {
43
+ "id": "thistle",
44
+ "index": 10
45
+ },
46
+ {
47
+ "id": "kincade",
48
+ "index": 11
49
+ },
50
+ {
51
+ "id": "poppy",
52
+ "index": 12
53
+ },
54
+ {
55
+ "id": "priya",
56
+ "index": 13
57
+ }
58
+ ]
coreai/coreai/duration_predictor.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ �0�M|�\��ճ!0R,zt��C��O�f��?f
coreai/coreai/duration_predictor.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fa30904d7f7cca5cabb4d5b32130522c7a7495ec43f89210154fca66b3ea3f66
3
+ size 10921604
coreai/coreai/duration_predictor.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "creationDate" : "20261104T160229Z",
3
+ "producer" : "coreai-core 1.0.0b3",
4
+ "assetVersion" : "2.0"
5
+ }
coreai/coreai/g2p.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ a,�{�G.�+��w��>��O����T]�q~6�!
coreai/coreai/g2p.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:612cb67bfc472eb22bceec7796ad3e97e54fb09aedcc545dd91f717e0f36b421
3
+ size 28377289
coreai/coreai/g2p.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "assetVersion" : "2.0",
3
+ "creationDate" : "20261104T160227Z",
4
+ "producer" : "coreai-core 1.0.0b3"
5
+ }
coreai/coreai/pitch_context_encoder.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ ��YG�2B��F��$tx��!}3`�5�
coreai/coreai/pitch_context_encoder.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bba55947ab3242f7b80146b3b6247478f6da217d3360b535dc0dfefaec5bd1a9
3
+ size 4700833
coreai/coreai/pitch_context_encoder.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "producer" : "coreai-core 1.0.0b3",
3
+ "assetVersion" : "2.0",
4
+ "creationDate" : "20261104T160231Z"
5
+ }
coreai/coreai/pitch_estimator.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ Tlo��g�7U6���I�mI� �������a��B
coreai/coreai/pitch_estimator.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:546c6f84e167b61b37553686e6f149aa6d49a60b888d99cd058aecc8619be142
3
+ size 25681787
coreai/coreai/pitch_estimator.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "creationDate" : "20261104T160239Z",
3
+ "producer" : "coreai-core 1.0.0b3",
4
+ "assetVersion" : "2.0"
5
+ }
coreai/coreai/vector_context_encoder.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ y7�{8��x�4�h����{�F��� �+�
coreai/coreai/vector_context_encoder.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:137937a27b388dd8788234a5681888ed81168e12de7be346c5d0de0c13db2b8e
3
+ size 8029278
coreai/coreai/vector_context_encoder.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "producer" : "coreai-core 1.0.0b3",
3
+ "creationDate" : "20261104T160230Z",
4
+ "assetVersion" : "2.0"
5
+ }
coreai/coreai/vector_estimator.aimodel/main.hash ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ fIi��r�bܑ�}�
2
+ #|�NaV��ZcV����M
coreai/coreai/vector_estimator.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:664969ffee72ee62dc91897ddb0a237c984e6156b8ed9d5a6356ae868914894d
3
+ size 63194410
coreai/coreai/vector_estimator.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "producer" : "coreai-core 1.0.0b3",
3
+ "assetVersion" : "2.0",
4
+ "creationDate" : "20261104T160253Z"
5
+ }
coreai/coreai/vocoder.aimodel/main.hash ADDED
@@ -0,0 +1 @@
 
 
1
+ b�(�Sj|=�ׅbY�����m�4���i���|�
coreai/coreai/vocoder.aimodel/main.mlirb ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62bd2889536a7c3de0d7856259c0a2f5d2cb7f6d8834fce0fb6994aea87c02da
3
+ size 43814084
coreai/coreai/vocoder.aimodel/metadata.json ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {
2
+ "producer" : "coreai-core 1.0.0b3",
3
+ "creationDate" : "20261104T160301Z",
4
+ "assetVersion" : "2.0"
5
+ }
coreai/manifest.json ADDED
@@ -0,0 +1,1483 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architecture": "scyllasband-duration-flow",
3
+ "assets": {
4
+ "fixed_delivery_references": "assets/fixed_references.json",
5
+ "g2p_config": "assets/g2p/export_config.json",
6
+ "g2p_language_map": "assets/g2p/language_map.json",
7
+ "g2p_normalization": "assets/g2p/normalization.json",
8
+ "g2p_phoneme_dict": "assets/g2p/phoneme_dict.json",
9
+ "g2p_pronunciation_overrides": "assets/g2p/pronunciation_overrides.json",
10
+ "g2p_tokenizer": "assets/g2p/tokenizer.json",
11
+ "language_index": "assets/languages.json",
12
+ "phone_vocab": "assets/phone_vocab.json",
13
+ "voice_index": "assets/voices.json"
14
+ },
15
+ "audio": {
16
+ "hop_length": 256,
17
+ "latent_dim": 24,
18
+ "latent_hop_length": 512,
19
+ "n_mels": 100,
20
+ "sample_rate": 24000
21
+ },
22
+ "components": {
23
+ "g2p_64": {
24
+ "artifacts": {
25
+ "coreai": {
26
+ "format": "coreai",
27
+ "path": "coreai/g2p.aimodel",
28
+ "function": "g2p_64"
29
+ }
30
+ },
31
+ "inputs": [
32
+ "text"
33
+ ],
34
+ "outputs": [
35
+ "logits"
36
+ ],
37
+ "required": false
38
+ },
39
+ "g2p_128": {
40
+ "artifacts": {
41
+ "coreai": {
42
+ "format": "coreai",
43
+ "path": "coreai/g2p.aimodel",
44
+ "function": "g2p_128"
45
+ }
46
+ },
47
+ "inputs": [
48
+ "text"
49
+ ],
50
+ "outputs": [
51
+ "logits"
52
+ ],
53
+ "required": false
54
+ },
55
+ "g2p_256": {
56
+ "artifacts": {
57
+ "coreai": {
58
+ "format": "coreai",
59
+ "path": "coreai/g2p.aimodel",
60
+ "function": "g2p_256"
61
+ }
62
+ },
63
+ "inputs": [
64
+ "text"
65
+ ],
66
+ "outputs": [
67
+ "logits"
68
+ ],
69
+ "required": false
70
+ },
71
+ "g2p": {
72
+ "artifacts": {
73
+ "coreai": {
74
+ "format": "coreai",
75
+ "path": "coreai/g2p.aimodel",
76
+ "function": "g2p"
77
+ }
78
+ },
79
+ "inputs": [
80
+ "text"
81
+ ],
82
+ "outputs": [
83
+ "logits"
84
+ ],
85
+ "required": true
86
+ },
87
+ "duration_predictor": {
88
+ "artifacts": {
89
+ "coreai": {
90
+ "format": "coreai",
91
+ "path": "coreai/duration_predictor.aimodel",
92
+ "function": "duration_predictor"
93
+ }
94
+ },
95
+ "inputs": [
96
+ "span_phone_ids",
97
+ "span_segment_ids",
98
+ "span_mask",
99
+ "voice_id",
100
+ "language_id",
101
+ "delivery_values",
102
+ "delivery_present"
103
+ ],
104
+ "outputs": [
105
+ "durations"
106
+ ],
107
+ "required": true
108
+ },
109
+ "vector_context_encoder": {
110
+ "artifacts": {
111
+ "coreai": {
112
+ "format": "coreai",
113
+ "path": "coreai/vector_context_encoder.aimodel",
114
+ "function": "vector_context_encoder"
115
+ }
116
+ },
117
+ "inputs": [
118
+ "span_context_phone_ids",
119
+ "span_context_segment_ids",
120
+ "span_context_mask"
121
+ ],
122
+ "outputs": [
123
+ "span_context_hidden"
124
+ ],
125
+ "required": true
126
+ },
127
+ "pitch_context_encoder": {
128
+ "artifacts": {
129
+ "coreai": {
130
+ "format": "coreai",
131
+ "path": "coreai/pitch_context_encoder.aimodel",
132
+ "function": "pitch_context_encoder"
133
+ }
134
+ },
135
+ "inputs": [
136
+ "span_context_phone_ids",
137
+ "span_context_segment_ids",
138
+ "span_context_mask"
139
+ ],
140
+ "outputs": [
141
+ "span_context_hidden"
142
+ ],
143
+ "required": true
144
+ },
145
+ "vector_estimator_128": {
146
+ "artifacts": {
147
+ "coreai": {
148
+ "format": "coreai",
149
+ "path": "coreai/vector_estimator.aimodel",
150
+ "function": "vector_estimator_128"
151
+ }
152
+ },
153
+ "inputs": [
154
+ "noise",
155
+ "time",
156
+ "expanded_phone_ids",
157
+ "voice_id",
158
+ "language_id",
159
+ "delivery_values",
160
+ "delivery_present",
161
+ "latent_mask",
162
+ "span_context_hidden",
163
+ "prefix_latents",
164
+ "prefix_mask",
165
+ "expanded_boundary_event_ids",
166
+ "expanded_modifier_event_ids",
167
+ "expanded_phone_phase",
168
+ "expanded_phone_log_duration",
169
+ "expanded_sentence_type_ids",
170
+ "expanded_melody"
171
+ ],
172
+ "outputs": [
173
+ "velocity"
174
+ ],
175
+ "required": false
176
+ },
177
+ "pitch_estimator_128": {
178
+ "artifacts": {
179
+ "coreai": {
180
+ "format": "coreai",
181
+ "path": "coreai/pitch_estimator.aimodel",
182
+ "function": "pitch_estimator_128"
183
+ }
184
+ },
185
+ "inputs": [
186
+ "noise",
187
+ "time",
188
+ "expanded_phone_ids",
189
+ "voice_id",
190
+ "language_id",
191
+ "delivery_values",
192
+ "delivery_present",
193
+ "latent_mask",
194
+ "span_context_hidden",
195
+ "expanded_boundary_event_ids",
196
+ "expanded_modifier_event_ids",
197
+ "expanded_phone_phase",
198
+ "expanded_phone_log_duration",
199
+ "expanded_sentence_type_ids"
200
+ ],
201
+ "outputs": [
202
+ "velocity"
203
+ ],
204
+ "required": false
205
+ },
206
+ "vocoder_128": {
207
+ "artifacts": {
208
+ "coreai": {
209
+ "format": "coreai",
210
+ "path": "coreai/vocoder.aimodel",
211
+ "function": "vocoder_128"
212
+ }
213
+ },
214
+ "inputs": [
215
+ "latents",
216
+ "latent_mask",
217
+ "voice_id",
218
+ "language_id",
219
+ "emotion_id"
220
+ ],
221
+ "outputs": [
222
+ "audio"
223
+ ],
224
+ "required": false
225
+ },
226
+ "vector_estimator_256": {
227
+ "artifacts": {
228
+ "coreai": {
229
+ "format": "coreai",
230
+ "path": "coreai/vector_estimator.aimodel",
231
+ "function": "vector_estimator_256"
232
+ }
233
+ },
234
+ "inputs": [
235
+ "noise",
236
+ "time",
237
+ "expanded_phone_ids",
238
+ "voice_id",
239
+ "language_id",
240
+ "delivery_values",
241
+ "delivery_present",
242
+ "latent_mask",
243
+ "span_context_hidden",
244
+ "prefix_latents",
245
+ "prefix_mask",
246
+ "expanded_boundary_event_ids",
247
+ "expanded_modifier_event_ids",
248
+ "expanded_phone_phase",
249
+ "expanded_phone_log_duration",
250
+ "expanded_sentence_type_ids",
251
+ "expanded_melody"
252
+ ],
253
+ "outputs": [
254
+ "velocity"
255
+ ],
256
+ "required": false
257
+ },
258
+ "pitch_estimator_256": {
259
+ "artifacts": {
260
+ "coreai": {
261
+ "format": "coreai",
262
+ "path": "coreai/pitch_estimator.aimodel",
263
+ "function": "pitch_estimator_256"
264
+ }
265
+ },
266
+ "inputs": [
267
+ "noise",
268
+ "time",
269
+ "expanded_phone_ids",
270
+ "voice_id",
271
+ "language_id",
272
+ "delivery_values",
273
+ "delivery_present",
274
+ "latent_mask",
275
+ "span_context_hidden",
276
+ "expanded_boundary_event_ids",
277
+ "expanded_modifier_event_ids",
278
+ "expanded_phone_phase",
279
+ "expanded_phone_log_duration",
280
+ "expanded_sentence_type_ids"
281
+ ],
282
+ "outputs": [
283
+ "velocity"
284
+ ],
285
+ "required": false
286
+ },
287
+ "vocoder_256": {
288
+ "artifacts": {
289
+ "coreai": {
290
+ "format": "coreai",
291
+ "path": "coreai/vocoder.aimodel",
292
+ "function": "vocoder_256"
293
+ }
294
+ },
295
+ "inputs": [
296
+ "latents",
297
+ "latent_mask",
298
+ "voice_id",
299
+ "language_id",
300
+ "emotion_id"
301
+ ],
302
+ "outputs": [
303
+ "audio"
304
+ ],
305
+ "required": false
306
+ },
307
+ "vector_estimator_384": {
308
+ "artifacts": {
309
+ "coreai": {
310
+ "format": "coreai",
311
+ "path": "coreai/vector_estimator.aimodel",
312
+ "function": "vector_estimator_384"
313
+ }
314
+ },
315
+ "inputs": [
316
+ "noise",
317
+ "time",
318
+ "expanded_phone_ids",
319
+ "voice_id",
320
+ "language_id",
321
+ "delivery_values",
322
+ "delivery_present",
323
+ "latent_mask",
324
+ "span_context_hidden",
325
+ "prefix_latents",
326
+ "prefix_mask",
327
+ "expanded_boundary_event_ids",
328
+ "expanded_modifier_event_ids",
329
+ "expanded_phone_phase",
330
+ "expanded_phone_log_duration",
331
+ "expanded_sentence_type_ids",
332
+ "expanded_melody"
333
+ ],
334
+ "outputs": [
335
+ "velocity"
336
+ ],
337
+ "required": false
338
+ },
339
+ "pitch_estimator_384": {
340
+ "artifacts": {
341
+ "coreai": {
342
+ "format": "coreai",
343
+ "path": "coreai/pitch_estimator.aimodel",
344
+ "function": "pitch_estimator_384"
345
+ }
346
+ },
347
+ "inputs": [
348
+ "noise",
349
+ "time",
350
+ "expanded_phone_ids",
351
+ "voice_id",
352
+ "language_id",
353
+ "delivery_values",
354
+ "delivery_present",
355
+ "latent_mask",
356
+ "span_context_hidden",
357
+ "expanded_boundary_event_ids",
358
+ "expanded_modifier_event_ids",
359
+ "expanded_phone_phase",
360
+ "expanded_phone_log_duration",
361
+ "expanded_sentence_type_ids"
362
+ ],
363
+ "outputs": [
364
+ "velocity"
365
+ ],
366
+ "required": false
367
+ },
368
+ "vocoder_384": {
369
+ "artifacts": {
370
+ "coreai": {
371
+ "format": "coreai",
372
+ "path": "coreai/vocoder.aimodel",
373
+ "function": "vocoder_384"
374
+ }
375
+ },
376
+ "inputs": [
377
+ "latents",
378
+ "latent_mask",
379
+ "voice_id",
380
+ "language_id",
381
+ "emotion_id"
382
+ ],
383
+ "outputs": [
384
+ "audio"
385
+ ],
386
+ "required": false
387
+ },
388
+ "vector_estimator_512": {
389
+ "artifacts": {
390
+ "coreai": {
391
+ "format": "coreai",
392
+ "path": "coreai/vector_estimator.aimodel",
393
+ "function": "vector_estimator_512"
394
+ }
395
+ },
396
+ "inputs": [
397
+ "noise",
398
+ "time",
399
+ "expanded_phone_ids",
400
+ "voice_id",
401
+ "language_id",
402
+ "delivery_values",
403
+ "delivery_present",
404
+ "latent_mask",
405
+ "span_context_hidden",
406
+ "prefix_latents",
407
+ "prefix_mask",
408
+ "expanded_boundary_event_ids",
409
+ "expanded_modifier_event_ids",
410
+ "expanded_phone_phase",
411
+ "expanded_phone_log_duration",
412
+ "expanded_sentence_type_ids",
413
+ "expanded_melody"
414
+ ],
415
+ "outputs": [
416
+ "velocity"
417
+ ],
418
+ "required": false
419
+ },
420
+ "pitch_estimator_512": {
421
+ "artifacts": {
422
+ "coreai": {
423
+ "format": "coreai",
424
+ "path": "coreai/pitch_estimator.aimodel",
425
+ "function": "pitch_estimator_512"
426
+ }
427
+ },
428
+ "inputs": [
429
+ "noise",
430
+ "time",
431
+ "expanded_phone_ids",
432
+ "voice_id",
433
+ "language_id",
434
+ "delivery_values",
435
+ "delivery_present",
436
+ "latent_mask",
437
+ "span_context_hidden",
438
+ "expanded_boundary_event_ids",
439
+ "expanded_modifier_event_ids",
440
+ "expanded_phone_phase",
441
+ "expanded_phone_log_duration",
442
+ "expanded_sentence_type_ids"
443
+ ],
444
+ "outputs": [
445
+ "velocity"
446
+ ],
447
+ "required": false
448
+ },
449
+ "vocoder_512": {
450
+ "artifacts": {
451
+ "coreai": {
452
+ "format": "coreai",
453
+ "path": "coreai/vocoder.aimodel",
454
+ "function": "vocoder_512"
455
+ }
456
+ },
457
+ "inputs": [
458
+ "latents",
459
+ "latent_mask",
460
+ "voice_id",
461
+ "language_id",
462
+ "emotion_id"
463
+ ],
464
+ "outputs": [
465
+ "audio"
466
+ ],
467
+ "required": false
468
+ },
469
+ "vector_estimator_768": {
470
+ "artifacts": {
471
+ "coreai": {
472
+ "format": "coreai",
473
+ "path": "coreai/vector_estimator.aimodel",
474
+ "function": "vector_estimator_768"
475
+ }
476
+ },
477
+ "inputs": [
478
+ "noise",
479
+ "time",
480
+ "expanded_phone_ids",
481
+ "voice_id",
482
+ "language_id",
483
+ "delivery_values",
484
+ "delivery_present",
485
+ "latent_mask",
486
+ "span_context_hidden",
487
+ "prefix_latents",
488
+ "prefix_mask",
489
+ "expanded_boundary_event_ids",
490
+ "expanded_modifier_event_ids",
491
+ "expanded_phone_phase",
492
+ "expanded_phone_log_duration",
493
+ "expanded_sentence_type_ids",
494
+ "expanded_melody"
495
+ ],
496
+ "outputs": [
497
+ "velocity"
498
+ ],
499
+ "required": false
500
+ },
501
+ "pitch_estimator_768": {
502
+ "artifacts": {
503
+ "coreai": {
504
+ "format": "coreai",
505
+ "path": "coreai/pitch_estimator.aimodel",
506
+ "function": "pitch_estimator_768"
507
+ }
508
+ },
509
+ "inputs": [
510
+ "noise",
511
+ "time",
512
+ "expanded_phone_ids",
513
+ "voice_id",
514
+ "language_id",
515
+ "delivery_values",
516
+ "delivery_present",
517
+ "latent_mask",
518
+ "span_context_hidden",
519
+ "expanded_boundary_event_ids",
520
+ "expanded_modifier_event_ids",
521
+ "expanded_phone_phase",
522
+ "expanded_phone_log_duration",
523
+ "expanded_sentence_type_ids"
524
+ ],
525
+ "outputs": [
526
+ "velocity"
527
+ ],
528
+ "required": false
529
+ },
530
+ "vocoder_768": {
531
+ "artifacts": {
532
+ "coreai": {
533
+ "format": "coreai",
534
+ "path": "coreai/vocoder.aimodel",
535
+ "function": "vocoder_768"
536
+ }
537
+ },
538
+ "inputs": [
539
+ "latents",
540
+ "latent_mask",
541
+ "voice_id",
542
+ "language_id",
543
+ "emotion_id"
544
+ ],
545
+ "outputs": [
546
+ "audio"
547
+ ],
548
+ "required": false
549
+ },
550
+ "vector_estimator": {
551
+ "artifacts": {
552
+ "coreai": {
553
+ "format": "coreai",
554
+ "path": "coreai/vector_estimator.aimodel",
555
+ "function": "vector_estimator"
556
+ }
557
+ },
558
+ "inputs": [
559
+ "noise",
560
+ "time",
561
+ "expanded_phone_ids",
562
+ "voice_id",
563
+ "language_id",
564
+ "delivery_values",
565
+ "delivery_present",
566
+ "latent_mask",
567
+ "span_context_hidden",
568
+ "prefix_latents",
569
+ "prefix_mask",
570
+ "expanded_boundary_event_ids",
571
+ "expanded_modifier_event_ids",
572
+ "expanded_phone_phase",
573
+ "expanded_phone_log_duration",
574
+ "expanded_sentence_type_ids",
575
+ "expanded_melody"
576
+ ],
577
+ "outputs": [
578
+ "velocity"
579
+ ],
580
+ "required": true
581
+ },
582
+ "pitch_estimator": {
583
+ "artifacts": {
584
+ "coreai": {
585
+ "format": "coreai",
586
+ "path": "coreai/pitch_estimator.aimodel",
587
+ "function": "pitch_estimator"
588
+ }
589
+ },
590
+ "inputs": [
591
+ "noise",
592
+ "time",
593
+ "expanded_phone_ids",
594
+ "voice_id",
595
+ "language_id",
596
+ "delivery_values",
597
+ "delivery_present",
598
+ "latent_mask",
599
+ "span_context_hidden",
600
+ "expanded_boundary_event_ids",
601
+ "expanded_modifier_event_ids",
602
+ "expanded_phone_phase",
603
+ "expanded_phone_log_duration",
604
+ "expanded_sentence_type_ids"
605
+ ],
606
+ "outputs": [
607
+ "velocity"
608
+ ],
609
+ "required": true
610
+ },
611
+ "vocoder": {
612
+ "artifacts": {
613
+ "coreai": {
614
+ "format": "coreai",
615
+ "path": "coreai/vocoder.aimodel",
616
+ "function": "vocoder"
617
+ }
618
+ },
619
+ "inputs": [
620
+ "latents",
621
+ "latent_mask",
622
+ "voice_id",
623
+ "language_id",
624
+ "emotion_id"
625
+ ],
626
+ "outputs": [
627
+ "audio"
628
+ ],
629
+ "required": true
630
+ }
631
+ },
632
+ "contract_version": "1.0.0",
633
+ "controls": {
634
+ "chunk_context": {
635
+ "artificial_continuation_boundaries": [
636
+ "chunk_continue"
637
+ ],
638
+ "artificial_continuation_phone": null,
639
+ "boundary_after_values": [
640
+ "sentence_end",
641
+ "paragraph_end",
642
+ "clause_continue",
643
+ "chunk_continue"
644
+ ],
645
+ "boundary_before_values": [
646
+ "sentence_start",
647
+ "paragraph_start",
648
+ "clause_continue",
649
+ "chunk_continue"
650
+ ],
651
+ "clause_continuation_fallback_phone": "<pause_comma>",
652
+ "conditioning": "boundary_embeddings",
653
+ "context_phone_tokens": {
654
+ "continuation": "<ctx_continuation>",
655
+ "sentence_end": "<ctx_sentence_end>",
656
+ "sentence_start": "<ctx_sentence_start>"
657
+ },
658
+ "default_boundary_after": "sentence_end",
659
+ "default_boundary_before": "sentence_start",
660
+ "emit_context_phone_tokens": false,
661
+ "linguistic_continuation_boundaries": [
662
+ "clause_continue"
663
+ ],
664
+ "schema_version": 2
665
+ },
666
+ "chunking": {
667
+ "shared_split_phone": "<sil>",
668
+ "split_after": [
669
+ "<end_stmt>",
670
+ "<end_question>",
671
+ "<end_exclaim>",
672
+ "<ellipsis>"
673
+ ],
674
+ "unit": "sentence"
675
+ },
676
+ "decoding": {
677
+ "context_frames": 48,
678
+ "mode": "overlap"
679
+ },
680
+ "delivery": {
681
+ "defaults": {
682
+ "ariadne": {
683
+ "intensity": 0.0,
684
+ "pace": 0.0,
685
+ "pitch": 0.0
686
+ },
687
+ "felix": {
688
+ "intensity": 0.0,
689
+ "pace": 0.0,
690
+ "pitch": 0.0
691
+ },
692
+ "gwen": {
693
+ "intensity": 0.0,
694
+ "pace": 0.0,
695
+ "pitch": 0.0
696
+ },
697
+ "ink": {
698
+ "intensity": 0.0,
699
+ "pace": 0.0,
700
+ "pitch": 0.0
701
+ },
702
+ "kincade": {
703
+ "intensity": 0.0,
704
+ "pace": 0.0,
705
+ "pitch": 0.0
706
+ },
707
+ "max": {
708
+ "intensity": 0.0,
709
+ "pace": 0.0,
710
+ "pitch": 0.0
711
+ },
712
+ "orpheus": {
713
+ "intensity": 0.0,
714
+ "pace": 0.0,
715
+ "pitch": 0.0
716
+ },
717
+ "poppy": {
718
+ "intensity": 0.0,
719
+ "pace": 0.0,
720
+ "pitch": 0.0
721
+ },
722
+ "priya": {
723
+ "intensity": 0.0,
724
+ "pace": 0.0,
725
+ "pitch": 0.0
726
+ },
727
+ "rex": {
728
+ "intensity": 0.0,
729
+ "pace": 0.0,
730
+ "pitch": 0.0
731
+ },
732
+ "scylla": {
733
+ "intensity": 0.0,
734
+ "pace": 0.0,
735
+ "pitch": 0.0
736
+ },
737
+ "stone": {
738
+ "intensity": 0.0,
739
+ "pace": 0.0,
740
+ "pitch": 0.0
741
+ },
742
+ "thistle": {
743
+ "intensity": 0.0,
744
+ "pace": 0.0,
745
+ "pitch": 0.0
746
+ },
747
+ "tuesday": {
748
+ "intensity": 0.0,
749
+ "pace": 0.0,
750
+ "pitch": 0.0
751
+ }
752
+ },
753
+ "graph_slots": 5,
754
+ "min_spoken_phones": 4,
755
+ "ranges": {
756
+ "intensity": [
757
+ -2.0,
758
+ 2.0
759
+ ],
760
+ "pace": [
761
+ -2.0,
762
+ 2.0
763
+ ],
764
+ "pitch": [
765
+ -2.0,
766
+ 2.0
767
+ ]
768
+ },
769
+ "release_id": "v3-20261104",
770
+ "rule": "value = clip(z, -z_limit, z_limit) / z_scale, present; both slots absent on a target with fewer than min_spoken_phones spoken phones; slots 2-4 always absent",
771
+ "schema": "measured_controls_v1",
772
+ "slots": [
773
+ "intensity",
774
+ "pace"
775
+ ],
776
+ "units": {
777
+ "intensity": "sd",
778
+ "pace": "sd",
779
+ "pitch": "semitones"
780
+ },
781
+ "voice_locales": {
782
+ "ariadne": [
783
+ "de",
784
+ "en_us",
785
+ "es",
786
+ "fr",
787
+ "it",
788
+ "vi"
789
+ ],
790
+ "felix": [
791
+ "de",
792
+ "en_us",
793
+ "es",
794
+ "fr",
795
+ "it",
796
+ "vi"
797
+ ],
798
+ "gwen": [
799
+ "de",
800
+ "en_us",
801
+ "es",
802
+ "fr",
803
+ "it",
804
+ "vi"
805
+ ],
806
+ "ink": [
807
+ "de",
808
+ "en_gb",
809
+ "es",
810
+ "fr",
811
+ "it",
812
+ "vi"
813
+ ],
814
+ "kincade": [
815
+ "en_us",
816
+ "es",
817
+ "de",
818
+ "fr",
819
+ "it",
820
+ "vi"
821
+ ],
822
+ "max": [
823
+ "de",
824
+ "en_us",
825
+ "es",
826
+ "fr",
827
+ "it",
828
+ "vi"
829
+ ],
830
+ "orpheus": [
831
+ "de",
832
+ "en_gb",
833
+ "es",
834
+ "fr",
835
+ "it",
836
+ "vi"
837
+ ],
838
+ "poppy": [
839
+ "en_us",
840
+ "es",
841
+ "de",
842
+ "fr",
843
+ "it",
844
+ "vi"
845
+ ],
846
+ "priya": [
847
+ "en_us",
848
+ "es",
849
+ "de",
850
+ "fr",
851
+ "it",
852
+ "vi"
853
+ ],
854
+ "rex": [
855
+ "de",
856
+ "en_us",
857
+ "es",
858
+ "fr",
859
+ "it",
860
+ "vi"
861
+ ],
862
+ "scylla": [
863
+ "de",
864
+ "en_us",
865
+ "es",
866
+ "fr",
867
+ "it",
868
+ "vi"
869
+ ],
870
+ "stone": [
871
+ "de",
872
+ "en_us",
873
+ "es",
874
+ "fr",
875
+ "it",
876
+ "vi"
877
+ ],
878
+ "thistle": [
879
+ "en_us",
880
+ "es",
881
+ "de",
882
+ "fr",
883
+ "it",
884
+ "vi"
885
+ ],
886
+ "tuesday": [
887
+ "de",
888
+ "en_gb",
889
+ "es",
890
+ "fr",
891
+ "it",
892
+ "vi"
893
+ ]
894
+ },
895
+ "z_limit": 2.5,
896
+ "z_scale": 5.0
897
+ },
898
+ "durations": {
899
+ "rounding": "half_even",
900
+ "silence_min_frames": 0,
901
+ "speech_min_frames": 1,
902
+ "zero_frame_phones": "punctuation_and_modifiers"
903
+ },
904
+ "emotions": {
905
+ "enabled": false
906
+ },
907
+ "energy": {
908
+ "default": 1.0,
909
+ "max": 1.5,
910
+ "min": 0.5
911
+ },
912
+ "fixed_shapes": {
913
+ "g2p_text_tokens": 512,
914
+ "latent_frames": 1024,
915
+ "phone_frames": 512
916
+ },
917
+ "g2p_buckets": [
918
+ {
919
+ "text_tokens": 64,
920
+ "component": "g2p_64"
921
+ },
922
+ {
923
+ "text_tokens": 128,
924
+ "component": "g2p_128"
925
+ },
926
+ {
927
+ "text_tokens": 256,
928
+ "component": "g2p_256"
929
+ },
930
+ {
931
+ "text_tokens": 512,
932
+ "component": "g2p"
933
+ }
934
+ ],
935
+ "graph_input_contract": "scyllasband_measured_controls_v3",
936
+ "pace_pauses": {
937
+ "per_sd": -0.13,
938
+ "phone": "<sil>",
939
+ "rule": "before rounding, each <sil> position's predicted frames are multiplied by exp(per_sd * pace)"
940
+ },
941
+ "pitch": {
942
+ "center": 10.0,
943
+ "channels": 2,
944
+ "context_encoder": "pitch_context_encoder",
945
+ "edges": {
946
+ "ariadne": [
947
+ 7.69,
948
+ 27.31
949
+ ],
950
+ "felix": [
951
+ -3.02,
952
+ 18.91
953
+ ],
954
+ "gwen": [
955
+ 4.84,
956
+ 25.68
957
+ ],
958
+ "ink": [
959
+ 3.83,
960
+ 24.19
961
+ ],
962
+ "kincade": [
963
+ -6.17,
964
+ 18.61
965
+ ],
966
+ "max": [
967
+ -9.9,
968
+ 17.2
969
+ ],
970
+ "orpheus": [
971
+ -4.35,
972
+ 18.42
973
+ ],
974
+ "poppy": [
975
+ 6.97,
976
+ 25.88
977
+ ],
978
+ "priya": [
979
+ 6.0,
980
+ 25.96
981
+ ],
982
+ "rex": [
983
+ -3.75,
984
+ 17.13
985
+ ],
986
+ "scylla": [
987
+ 8.47,
988
+ 28.12
989
+ ],
990
+ "stone": [
991
+ -7.92,
992
+ 17.8
993
+ ],
994
+ "thistle": [
995
+ 6.5,
996
+ 26.67
997
+ ],
998
+ "tuesday": [
999
+ 6.13,
1000
+ 24.86
1001
+ ]
1002
+ },
1003
+ "knee": 2.0,
1004
+ "melody_center": 16.65,
1005
+ "melody_scale": 24.0,
1006
+ "rule": "contour [2, T]: st = c0 * scale + center + shift; st soft-limited inside the voice's edges (knee = min(knee, (high - low) / 4): unchanged until knee semitones from an edge, then edge -/+ knee * tanh); p = clip((c1 + 1) / 2, 0, 1); voiced = p > voiced_threshold; melody [T, 3] = ((st - melody_center) / melody_scale * voiced, p, voiced)",
1007
+ "sampler": {
1008
+ "choices": [
1009
+ "euler",
1010
+ "heun"
1011
+ ],
1012
+ "default": "euler"
1013
+ },
1014
+ "scale": 6.0,
1015
+ "steps": {
1016
+ "default": 4,
1017
+ "max": 64,
1018
+ "min": 1
1019
+ },
1020
+ "voiced_threshold": 0.5
1021
+ },
1022
+ "prefix_conditioning": {
1023
+ "alignment": "right",
1024
+ "enabled": true,
1025
+ "max_frames": 96,
1026
+ "source": "previous_latents_tail"
1027
+ },
1028
+ "punctuation_silence": {
1029
+ "calibrated_floors": false,
1030
+ "floor_table_ms": {},
1031
+ "silence_phone": "<sil>",
1032
+ "target": "explicit_silence"
1033
+ },
1034
+ "reference_packs": {
1035
+ "enabled": false,
1036
+ "fixed_in_graph": true
1037
+ },
1038
+ "sampler": {
1039
+ "choices": [
1040
+ "euler",
1041
+ "heun"
1042
+ ],
1043
+ "default": "euler"
1044
+ },
1045
+ "span_conditioning": {
1046
+ "context_max_phones": 512,
1047
+ "context_phones_each_side": 180,
1048
+ "duration_inputs": [
1049
+ "span_phone_ids",
1050
+ "span_segment_ids",
1051
+ "span_mask"
1052
+ ],
1053
+ "enabled": true,
1054
+ "encoder_component": "vector_context_encoder",
1055
+ "encoder_inputs": [
1056
+ "span_context_phone_ids",
1057
+ "span_context_segment_ids",
1058
+ "span_context_mask"
1059
+ ],
1060
+ "exclude_context_phones": [
1061
+ "<sil>"
1062
+ ],
1063
+ "hidden_size": 512,
1064
+ "layout": "balanced_before_target_after",
1065
+ "pitch_hidden_size": 384,
1066
+ "scope": "passage",
1067
+ "segments": 3,
1068
+ "solver_input": "span_context_hidden"
1069
+ },
1070
+ "speed": {
1071
+ "default": 1.0,
1072
+ "max": 1.25,
1073
+ "min": 0.75
1074
+ },
1075
+ "steps": {
1076
+ "default": 4,
1077
+ "max": 64,
1078
+ "min": 1
1079
+ },
1080
+ "target_buckets": {
1081
+ "bucket_margin_frames": 0,
1082
+ "buckets": [
1083
+ {
1084
+ "default": false,
1085
+ "latent_frames": 128,
1086
+ "vector_estimator": "vector_estimator_128",
1087
+ "pitch_estimator": "pitch_estimator_128",
1088
+ "vocoder": "vocoder_128"
1089
+ },
1090
+ {
1091
+ "default": false,
1092
+ "latent_frames": 256,
1093
+ "vector_estimator": "vector_estimator_256",
1094
+ "pitch_estimator": "pitch_estimator_256",
1095
+ "vocoder": "vocoder_256"
1096
+ },
1097
+ {
1098
+ "default": false,
1099
+ "latent_frames": 384,
1100
+ "vector_estimator": "vector_estimator_384",
1101
+ "pitch_estimator": "pitch_estimator_384",
1102
+ "vocoder": "vocoder_384"
1103
+ },
1104
+ {
1105
+ "default": false,
1106
+ "latent_frames": 512,
1107
+ "vector_estimator": "vector_estimator_512",
1108
+ "pitch_estimator": "pitch_estimator_512",
1109
+ "vocoder": "vocoder_512"
1110
+ },
1111
+ {
1112
+ "default": false,
1113
+ "latent_frames": 768,
1114
+ "vector_estimator": "vector_estimator_768",
1115
+ "pitch_estimator": "pitch_estimator_768",
1116
+ "vocoder": "vocoder_768"
1117
+ },
1118
+ {
1119
+ "default": true,
1120
+ "latent_frames": 1024,
1121
+ "vector_estimator": "vector_estimator",
1122
+ "pitch_estimator": "pitch_estimator",
1123
+ "vocoder": "vocoder"
1124
+ }
1125
+ ],
1126
+ "default_latent_frames": 1024,
1127
+ "enabled": true,
1128
+ "policy": "smallest_fit"
1129
+ },
1130
+ "text_normalizer": {
1131
+ "prepared_contract_recorded": true,
1132
+ "sha256": "d8fd1b2415d0d100d1c70237a4469117ee6b73d0e4da703e4654c83192779cc4",
1133
+ "version": "scyllasband_spoken_text_v2_2026_08_05"
1134
+ },
1135
+ "timing_events": {
1136
+ "enabled": true,
1137
+ "inputs": [
1138
+ "expanded_boundary_event_ids",
1139
+ "expanded_modifier_event_ids",
1140
+ "expanded_phone_phase",
1141
+ "expanded_phone_log_duration",
1142
+ "expanded_sentence_type_ids"
1143
+ ],
1144
+ "modifier_bits": {
1145
+ "ˈ": 1,
1146
+ "ˌ": 2,
1147
+ "ː": 4,
1148
+ "̃": 8,
1149
+ "̩": 16,
1150
+ "̪": 32
1151
+ },
1152
+ "punctuation_owner": "following_silence_else_first_frame_of_next_phone",
1153
+ "sentence_type_count": 5,
1154
+ "sentence_type_lookahead": "following_context",
1155
+ "sentence_types": {
1156
+ "<ellipsis>": 4,
1157
+ "<end_exclaim>": 3,
1158
+ "<end_question>": 2,
1159
+ "<end_stmt>": 1
1160
+ },
1161
+ "word_start_event_id": 87,
1162
+ "word_start_owner": "first_phone_of_word_after_chunk_first_word"
1163
+ },
1164
+ "word_boundaries": {
1165
+ "duration_phone": "<sil>",
1166
+ "enabled": true,
1167
+ "g2p_output_symbol": " ",
1168
+ "zero_frames_allowed": true
1169
+ },
1170
+ "coreai": {
1171
+ "minimum_os": {
1172
+ "ios": "27.0",
1173
+ "macos": "27.0"
1174
+ },
1175
+ "io": {
1176
+ "ids": "int32",
1177
+ "masks": "int32",
1178
+ "values": "float32"
1179
+ },
1180
+ "assets": {
1181
+ "g2p": {
1182
+ "precision": "fp16",
1183
+ "weights": "w8",
1184
+ "bytes": 28377426
1185
+ },
1186
+ "duration_predictor": {
1187
+ "precision": "fp16",
1188
+ "weights": "w8",
1189
+ "bytes": 10921741
1190
+ },
1191
+ "vector_context_encoder": {
1192
+ "precision": "fp16",
1193
+ "weights": "w8",
1194
+ "bytes": 8029415
1195
+ },
1196
+ "pitch_context_encoder": {
1197
+ "precision": "fp16",
1198
+ "weights": "w8",
1199
+ "bytes": 4700970
1200
+ },
1201
+ "pitch_estimator": {
1202
+ "precision": "fp16",
1203
+ "weights": "w8",
1204
+ "bytes": 25681924
1205
+ },
1206
+ "vector_estimator": {
1207
+ "precision": "fp16",
1208
+ "weights": "w8",
1209
+ "bytes": 63194547
1210
+ },
1211
+ "vocoder": {
1212
+ "precision": "fp16",
1213
+ "weights": "w8",
1214
+ "bytes": 43814221
1215
+ }
1216
+ },
1217
+ "compute_units": "gpu",
1218
+ "neural_engine_assets": [
1219
+ "vector_estimator",
1220
+ "pitch_estimator"
1221
+ ]
1222
+ }
1223
+ },
1224
+ "dataset_summary": {
1225
+ "controls": "measured_controls_v1",
1226
+ "voice_locale_pairs": 84
1227
+ },
1228
+ "default_language": "en_us",
1229
+ "languages": [
1230
+ "en_us",
1231
+ "en_gb",
1232
+ "es",
1233
+ "it",
1234
+ "fr",
1235
+ "de",
1236
+ "vi"
1237
+ ],
1238
+ "model_name": "Scylla's Band",
1239
+ "model_version": "3",
1240
+ "preferred_backends": [
1241
+ "coreai"
1242
+ ],
1243
+ "voices": [
1244
+ {
1245
+ "default_language": "en_us",
1246
+ "embedding_path": null,
1247
+ "id": "ariadne",
1248
+ "languages": [
1249
+ "de",
1250
+ "en_us",
1251
+ "es",
1252
+ "fr",
1253
+ "it",
1254
+ "vi"
1255
+ ],
1256
+ "name": "ariadne",
1257
+ "styles": [
1258
+ "default"
1259
+ ]
1260
+ },
1261
+ {
1262
+ "default_language": "en_us",
1263
+ "embedding_path": null,
1264
+ "id": "felix",
1265
+ "languages": [
1266
+ "de",
1267
+ "en_us",
1268
+ "es",
1269
+ "fr",
1270
+ "it",
1271
+ "vi"
1272
+ ],
1273
+ "name": "felix",
1274
+ "styles": [
1275
+ "default"
1276
+ ]
1277
+ },
1278
+ {
1279
+ "default_language": "en_us",
1280
+ "embedding_path": null,
1281
+ "id": "gwen",
1282
+ "languages": [
1283
+ "de",
1284
+ "en_us",
1285
+ "es",
1286
+ "fr",
1287
+ "it",
1288
+ "vi"
1289
+ ],
1290
+ "name": "gwen",
1291
+ "styles": [
1292
+ "default"
1293
+ ]
1294
+ },
1295
+ {
1296
+ "default_language": "en_gb",
1297
+ "embedding_path": null,
1298
+ "id": "ink",
1299
+ "languages": [
1300
+ "de",
1301
+ "en_gb",
1302
+ "es",
1303
+ "fr",
1304
+ "it",
1305
+ "vi"
1306
+ ],
1307
+ "name": "ink",
1308
+ "styles": [
1309
+ "default"
1310
+ ]
1311
+ },
1312
+ {
1313
+ "default_language": "en_us",
1314
+ "embedding_path": null,
1315
+ "id": "max",
1316
+ "languages": [
1317
+ "de",
1318
+ "en_us",
1319
+ "es",
1320
+ "fr",
1321
+ "it",
1322
+ "vi"
1323
+ ],
1324
+ "name": "max",
1325
+ "styles": [
1326
+ "default"
1327
+ ]
1328
+ },
1329
+ {
1330
+ "default_language": "en_gb",
1331
+ "embedding_path": null,
1332
+ "id": "orpheus",
1333
+ "languages": [
1334
+ "de",
1335
+ "en_gb",
1336
+ "es",
1337
+ "fr",
1338
+ "it",
1339
+ "vi"
1340
+ ],
1341
+ "name": "orpheus",
1342
+ "styles": [
1343
+ "default"
1344
+ ]
1345
+ },
1346
+ {
1347
+ "default_language": "en_us",
1348
+ "embedding_path": null,
1349
+ "id": "rex",
1350
+ "languages": [
1351
+ "de",
1352
+ "en_us",
1353
+ "es",
1354
+ "fr",
1355
+ "it",
1356
+ "vi"
1357
+ ],
1358
+ "name": "rex",
1359
+ "styles": [
1360
+ "default"
1361
+ ]
1362
+ },
1363
+ {
1364
+ "default_language": "en_us",
1365
+ "embedding_path": null,
1366
+ "id": "scylla",
1367
+ "languages": [
1368
+ "de",
1369
+ "en_us",
1370
+ "es",
1371
+ "fr",
1372
+ "it",
1373
+ "vi"
1374
+ ],
1375
+ "name": "scylla",
1376
+ "styles": [
1377
+ "default"
1378
+ ]
1379
+ },
1380
+ {
1381
+ "default_language": "en_us",
1382
+ "embedding_path": null,
1383
+ "id": "stone",
1384
+ "languages": [
1385
+ "de",
1386
+ "en_us",
1387
+ "es",
1388
+ "fr",
1389
+ "it",
1390
+ "vi"
1391
+ ],
1392
+ "name": "stone",
1393
+ "styles": [
1394
+ "default"
1395
+ ]
1396
+ },
1397
+ {
1398
+ "default_language": "en_gb",
1399
+ "embedding_path": null,
1400
+ "id": "tuesday",
1401
+ "languages": [
1402
+ "de",
1403
+ "en_gb",
1404
+ "es",
1405
+ "fr",
1406
+ "it",
1407
+ "vi"
1408
+ ],
1409
+ "name": "tuesday",
1410
+ "styles": [
1411
+ "default"
1412
+ ]
1413
+ },
1414
+ {
1415
+ "default_language": "en_us",
1416
+ "embedding_path": null,
1417
+ "id": "thistle",
1418
+ "languages": [
1419
+ "en_us",
1420
+ "es",
1421
+ "de",
1422
+ "fr",
1423
+ "it",
1424
+ "vi"
1425
+ ],
1426
+ "name": "thistle",
1427
+ "styles": [
1428
+ "default"
1429
+ ]
1430
+ },
1431
+ {
1432
+ "default_language": "en_us",
1433
+ "embedding_path": null,
1434
+ "id": "kincade",
1435
+ "languages": [
1436
+ "en_us",
1437
+ "es",
1438
+ "de",
1439
+ "fr",
1440
+ "it",
1441
+ "vi"
1442
+ ],
1443
+ "name": "kincade",
1444
+ "styles": [
1445
+ "default"
1446
+ ]
1447
+ },
1448
+ {
1449
+ "default_language": "en_us",
1450
+ "embedding_path": null,
1451
+ "id": "poppy",
1452
+ "languages": [
1453
+ "en_us",
1454
+ "es",
1455
+ "de",
1456
+ "fr",
1457
+ "it",
1458
+ "vi"
1459
+ ],
1460
+ "name": "poppy",
1461
+ "styles": [
1462
+ "default"
1463
+ ]
1464
+ },
1465
+ {
1466
+ "default_language": "en_us",
1467
+ "embedding_path": null,
1468
+ "id": "priya",
1469
+ "languages": [
1470
+ "en_us",
1471
+ "es",
1472
+ "de",
1473
+ "fr",
1474
+ "it",
1475
+ "vi"
1476
+ ],
1477
+ "name": "priya",
1478
+ "styles": [
1479
+ "default"
1480
+ ]
1481
+ }
1482
+ ]
1483
+ }
coreml/assets/fixed_references.json ADDED
The diff for this file is too large to render. See raw diff
 
coreml/assets/g2p/export_config.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "batch_dynamic": true,
3
+ "format": "scyllasband_g2p",
4
+ "input_granularity": "phrase",
5
+ "languages": [
6
+ "de",
7
+ "en_gb",
8
+ "en_us",
9
+ "es",
10
+ "fr",
11
+ "it",
12
+ "vi"
13
+ ],
14
+ "model_type": "transformer",
15
+ "onnx_export_batch_size": 4,
16
+ "onnx_export_layout": "batch_first",
17
+ "onnx_export_max_batch_size": 256,
18
+ "onnx_exporter": "dynamo",
19
+ "onnx_fp32_model": "model.onnx",
20
+ "onnx_model": "model.onnx",
21
+ "onnx_padded_output": "ctc_blank",
22
+ "onnx_quantization": "none",
23
+ "onnx_quantized_parity_tolerance": 3.0,
24
+ "opset": 18,
25
+ "phoneme_token_mode": "char",
26
+ "phrase_rows_per_language": 200000,
27
+ "preserve_commas": true,
28
+ "punctuation_silence_target": "explicit_silence",
29
+ "release_role": "fp32_bundle_frontend",
30
+ "source_checkpoint": "output/deepphonemizer_multilingual/model_v10_7lang/checkpoints/model_step_250k.pt",
31
+ "training_mix": "phrase_word_chunked_clean",
32
+ "word_rows_per_language": 200000
33
+ }
coreml/assets/g2p/language_map.json ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "american": "en_us",
3
+ "american_english": "en_us",
4
+ "british": "en_gb",
5
+ "british_english": "en_gb",
6
+ "de": "de",
7
+ "de_de": "de",
8
+ "deu": "de",
9
+ "en": "en_us",
10
+ "en-gb": "en_gb",
11
+ "en-uk": "en_gb",
12
+ "en-us": "en_us",
13
+ "en_gb": "en_gb",
14
+ "en_uk": "en_gb",
15
+ "en_us": "en_us",
16
+ "eng": "en_us",
17
+ "eng_gb": "en_gb",
18
+ "eng_us": "en_us",
19
+ "english": "en_us",
20
+ "english_gb": "en_gb",
21
+ "english_uk": "en_gb",
22
+ "english_us": "en_us",
23
+ "es": "es",
24
+ "es_es": "es",
25
+ "es_mx": "es",
26
+ "fr": "fr",
27
+ "fr_fr": "fr",
28
+ "fra": "fr",
29
+ "fre": "fr",
30
+ "french": "fr",
31
+ "gb": "en_gb",
32
+ "ger": "de",
33
+ "german": "de",
34
+ "it": "it",
35
+ "it_it": "it",
36
+ "ita": "it",
37
+ "italian": "it",
38
+ "spa": "es",
39
+ "spanish": "es",
40
+ "uk": "en_gb",
41
+ "us": "en_us",
42
+ "vi": "vi",
43
+ "vi_hn": "vi",
44
+ "vi_vn": "vi",
45
+ "vie": "vi",
46
+ "vietnamese": "vi"
47
+ }
coreml/assets/g2p/normalization.json ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "contract": {
3
+ "colon_semicolon_classes": "preserved",
4
+ "dot_runs_two_or_more": "ellipsis_three_dots",
5
+ "homogeneous_terminal_runs": "single_mark",
6
+ "lexical_hyphen": "space_no_pause",
7
+ "mixed_question_exclamation": "leading_mark",
8
+ "non_acoustic_modifiers": [
9
+ "stress",
10
+ "length",
11
+ "unicode_combining"
12
+ ],
13
+ "ok_spoken_form": "okay",
14
+ "punctuation_silence_target": "explicit_silence",
15
+ "syntactic_dash": "em_dash_pause",
16
+ "training_phone_source": "espeak",
17
+ "version": "scyllasband_spoken_text_v2_2026_08_05"
18
+ },
19
+ "contract_sha256": "d8fd1b2415d0d100d1c70237a4469117ee6b73d0e4da703e4654c83192779cc4",
20
+ "contract_version": "scyllasband_spoken_text_v2_2026_08_05",
21
+ "expand_currency": true,
22
+ "expand_dates": true,
23
+ "expand_numbers": true,
24
+ "expand_ordinals": true,
25
+ "expand_percentages": true,
26
+ "expand_times": true,
27
+ "languages": [
28
+ "en",
29
+ "en_us",
30
+ "en_gb",
31
+ "es",
32
+ "it",
33
+ "fr",
34
+ "de",
35
+ "vi"
36
+ ],
37
+ "normalize_at_sign": true,
38
+ "normalize_punctuation": true,
39
+ "normalizer": "SpokenTextNormalizer"
40
+ }
coreml/assets/g2p/phoneme_dict.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "de": {},
3
+ "en_gb": {},
4
+ "en_us": {},
5
+ "es": {},
6
+ "fr": {},
7
+ "it": {},
8
+ "vi": {}
9
+ }
coreml/assets/g2p/pronunciation_overrides.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "format": "scyllasband_g2p_pronunciation_overrides_v1",
3
+ "languages": {}
4
+ }
coreml/assets/g2p/tokenizer.json ADDED
@@ -0,0 +1,517 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "format": "deepphonemizer_forward_ctc",
3
+ "languages": [
4
+ "de",
5
+ "en_gb",
6
+ "en_us",
7
+ "es",
8
+ "fr",
9
+ "it",
10
+ "vi"
11
+ ],
12
+ "char_repeats": 4,
13
+ "lowercase": true,
14
+ "phoneme_token_mode": "char",
15
+ "text_symbols": {
16
+ "_": 0,
17
+ "<de>": 1,
18
+ "<en_gb>": 2,
19
+ "<en_us>": 3,
20
+ "<es>": 4,
21
+ "<fr>": 5,
22
+ "<it>": 6,
23
+ "<vi>": 7,
24
+ "<end>": 8,
25
+ "\b": 9,
26
+ " ": 10,
27
+ "!": 11,
28
+ "\"": 12,
29
+ "#": 13,
30
+ "$": 14,
31
+ "%": 15,
32
+ "&": 16,
33
+ "'": 17,
34
+ "(": 18,
35
+ ")": 19,
36
+ "*": 20,
37
+ "+": 21,
38
+ ",": 22,
39
+ "-": 23,
40
+ ".": 24,
41
+ "/": 25,
42
+ "0": 26,
43
+ "1": 27,
44
+ "2": 28,
45
+ "3": 29,
46
+ "4": 30,
47
+ "5": 31,
48
+ "6": 32,
49
+ "7": 33,
50
+ "8": 34,
51
+ "9": 35,
52
+ ":": 36,
53
+ ";": 37,
54
+ "<": 38,
55
+ "=": 39,
56
+ ">": 40,
57
+ "?": 41,
58
+ "@": 42,
59
+ "A": 43,
60
+ "B": 44,
61
+ "C": 45,
62
+ "D": 46,
63
+ "E": 47,
64
+ "F": 48,
65
+ "G": 49,
66
+ "H": 50,
67
+ "I": 51,
68
+ "J": 52,
69
+ "K": 53,
70
+ "L": 54,
71
+ "M": 55,
72
+ "N": 56,
73
+ "O": 57,
74
+ "P": 58,
75
+ "Q": 59,
76
+ "R": 60,
77
+ "S": 61,
78
+ "T": 62,
79
+ "U": 63,
80
+ "V": 64,
81
+ "W": 65,
82
+ "X": 66,
83
+ "Y": 67,
84
+ "Z": 68,
85
+ "[": 69,
86
+ "\\": 70,
87
+ "]": 71,
88
+ "^": 72,
89
+ "`": 73,
90
+ "a": 74,
91
+ "b": 75,
92
+ "c": 76,
93
+ "d": 77,
94
+ "e": 78,
95
+ "f": 79,
96
+ "g": 80,
97
+ "h": 81,
98
+ "i": 82,
99
+ "j": 83,
100
+ "k": 84,
101
+ "l": 85,
102
+ "m": 86,
103
+ "n": 87,
104
+ "o": 88,
105
+ "p": 89,
106
+ "q": 90,
107
+ "r": 91,
108
+ "s": 92,
109
+ "t": 93,
110
+ "u": 94,
111
+ "v": 95,
112
+ "w": 96,
113
+ "x": 97,
114
+ "y": 98,
115
+ "z": 99,
116
+ "{": 100,
117
+ "|": 101,
118
+ "}": 102,
119
+ "~": 103,
120
+ "¡": 104,
121
+ "¢": 105,
122
+ "£": 106,
123
+ "¥": 107,
124
+ "§": 108,
125
+ "ª": 109,
126
+ "«": 110,
127
+ "­": 111,
128
+ "®": 112,
129
+ "°": 113,
130
+ "±": 114,
131
+ "²": 115,
132
+ "³": 116,
133
+ "´": 117,
134
+ "¶": 118,
135
+ "·": 119,
136
+ "¹": 120,
137
+ "º": 121,
138
+ "»": 122,
139
+ "¼": 123,
140
+ "½": 124,
141
+ "¾": 125,
142
+ "¿": 126,
143
+ "À": 127,
144
+ "Á": 128,
145
+ "Â": 129,
146
+ "Ä": 130,
147
+ "Å": 131,
148
+ "Æ": 132,
149
+ "Ç": 133,
150
+ "È": 134,
151
+ "É": 135,
152
+ "Ê": 136,
153
+ "Í": 137,
154
+ "Ï": 138,
155
+ "Ñ": 139,
156
+ "Ó": 140,
157
+ "Ô": 141,
158
+ "Ö": 142,
159
+ "×": 143,
160
+ "Ù": 144,
161
+ "Ú": 145,
162
+ "Ü": 146,
163
+ "Ý": 147,
164
+ "Þ": 148,
165
+ "ß": 149,
166
+ "à": 150,
167
+ "á": 151,
168
+ "â": 152,
169
+ "ã": 153,
170
+ "ä": 154,
171
+ "å": 155,
172
+ "æ": 156,
173
+ "ç": 157,
174
+ "è": 158,
175
+ "é": 159,
176
+ "ê": 160,
177
+ "ë": 161,
178
+ "ì": 162,
179
+ "í": 163,
180
+ "î": 164,
181
+ "ï": 165,
182
+ "ð": 166,
183
+ "ñ": 167,
184
+ "ò": 168,
185
+ "ó": 169,
186
+ "ô": 170,
187
+ "õ": 171,
188
+ "ö": 172,
189
+ "ø": 173,
190
+ "ù": 174,
191
+ "ú": 175,
192
+ "û": 176,
193
+ "ü": 177,
194
+ "ý": 178,
195
+ "þ": 179,
196
+ "ÿ": 180,
197
+ "ā": 181,
198
+ "Ă": 182,
199
+ "ă": 183,
200
+ "ą": 184,
201
+ "ć": 185,
202
+ "č": 186,
203
+ "Đ": 187,
204
+ "đ": 188,
205
+ "ē": 189,
206
+ "ė": 190,
207
+ "ę": 191,
208
+ "ğ": 192,
209
+ "ġ": 193,
210
+ "ĩ": 194,
211
+ "ī": 195,
212
+ "ļ": 196,
213
+ "ł": 197,
214
+ "ń": 198,
215
+ "Ō": 199,
216
+ "ō": 200,
217
+ "Œ": 201,
218
+ "œ": 202,
219
+ "ŕ": 203,
220
+ "ś": 204,
221
+ "ş": 205,
222
+ "š": 206,
223
+ "ŧ": 207,
224
+ "ũ": 208,
225
+ "ū": 209,
226
+ "ŭ": 210,
227
+ "ů": 211,
228
+ "ű": 212,
229
+ "ż": 213,
230
+ "Ž": 214,
231
+ "ž": 215,
232
+ "ſ": 216,
233
+ "Ơ": 217,
234
+ "ơ": 218,
235
+ "Ư": 219,
236
+ "ư": 220,
237
+ "ǁ": 221,
238
+ "ǂ": 222,
239
+ "ǐ": 223,
240
+ "ǘ": 224,
241
+ "ǧ": 225,
242
+ "ș": 226,
243
+ "Ȝ": 227,
244
+ "ȝ": 228,
245
+ "ȟ": 229,
246
+ "ɛ": 230,
247
+ "ʰ": 231,
248
+ "ˆ": 232,
249
+ "ˈ": 233,
250
+ "ˢ": 234,
251
+ "̃": 235,
252
+ "̍": 236,
253
+ "̣": 237,
254
+ "̱": 238,
255
+ "͈": 239,
256
+ "α": 240,
257
+ "β": 241,
258
+ "γ": 242,
259
+ "δ": 243,
260
+ "ε": 244,
261
+ "η": 245,
262
+ "κ": 246,
263
+ "μ": 247,
264
+ "ξ": 248,
265
+ "π": 249,
266
+ "ς": 250,
267
+ "σ": 251,
268
+ "χ": 252,
269
+ "ψ": 253,
270
+ "а": 254,
271
+ "д": 255,
272
+ "е": 256,
273
+ "з": 257,
274
+ "и": 258,
275
+ "к": 259,
276
+ "л": 260,
277
+ "о": 261,
278
+ "п": 262,
279
+ "р": 263,
280
+ "с": 264,
281
+ "т": 265,
282
+ "ч": 266,
283
+ "ь": 267,
284
+ "ง": 268,
285
+ "ย": 269,
286
+ "า": 270,
287
+ "่": 271,
288
+ "ည": 272,
289
+ "လ": 273,
290
+ "း": 274,
291
+ "်": 275,
292
+ "ᵈ": 276,
293
+ "ᵗ": 277,
294
+ "ḅ": 278,
295
+ "ḍ": 279,
296
+ "ḳ": 280,
297
+ "ṅ": 281,
298
+ "Ṭ": 282,
299
+ "ẓ": 283,
300
+ "ạ": 284,
301
+ "Ả": 285,
302
+ "ả": 286,
303
+ "Ấ": 287,
304
+ "ấ": 288,
305
+ "ầ": 289,
306
+ "ẩ": 290,
307
+ "ẫ": 291,
308
+ "ậ": 292,
309
+ "ắ": 293,
310
+ "ằ": 294,
311
+ "ẳ": 295,
312
+ "ẵ": 296,
313
+ "ặ": 297,
314
+ "ẹ": 298,
315
+ "ẻ": 299,
316
+ "ẽ": 300,
317
+ "Ế": 301,
318
+ "ế": 302,
319
+ "Ề": 303,
320
+ "ề": 304,
321
+ "ể": 305,
322
+ "ễ": 306,
323
+ "ệ": 307,
324
+ "ỉ": 308,
325
+ "ị": 309,
326
+ "ọ": 310,
327
+ "ỏ": 311,
328
+ "Ố": 312,
329
+ "ố": 313,
330
+ "Ồ": 314,
331
+ "ồ": 315,
332
+ "Ổ": 316,
333
+ "ổ": 317,
334
+ "ỗ": 318,
335
+ "Ộ": 319,
336
+ "ộ": 320,
337
+ "ớ": 321,
338
+ "Ờ": 322,
339
+ "ờ": 323,
340
+ "Ở": 324,
341
+ "ở": 325,
342
+ "ỡ": 326,
343
+ "ợ": 327,
344
+ "ụ": 328,
345
+ "ủ": 329,
346
+ "Ứ": 330,
347
+ "ứ": 331,
348
+ "Ừ": 332,
349
+ "ừ": 333,
350
+ "ử": 334,
351
+ "ữ": 335,
352
+ "ự": 336,
353
+ "ỳ": 337,
354
+ "ỵ": 338,
355
+ "Ỷ": 339,
356
+ "ỷ": 340,
357
+ "ỹ": 341,
358
+ "ὖ": 342,
359
+ "ὼ": 343,
360
+ "ῆ": 344,
361
+ "​": 345,
362
+ "‐": 346,
363
+ "—": 347,
364
+ "―": 348,
365
+ "†": 349,
366
+ "•": 350,
367
+ "…": 351,
368
+ "‰": 352,
369
+ "′": 353,
370
+ "″": 354,
371
+ "⁄": 355,
372
+ "⁷": 356,
373
+ "⁻": 357,
374
+ "₀": 358,
375
+ "₁": 359,
376
+ "₂": 360,
377
+ "₃": 361,
378
+ "₄": 362,
379
+ "₅": 363,
380
+ "₆": 364,
381
+ "₇": 365,
382
+ "₈": 366,
383
+ "€": 367,
384
+ "₹": 368,
385
+ "⅓": 369,
386
+ "⅔": 370,
387
+ "⅖": 371,
388
+ "⅙": 372,
389
+ "⅛": 373,
390
+ "⅝": 374,
391
+ "√": 375,
392
+ "│": 376,
393
+ "♦": 377,
394
+ "⚓": 378,
395
+ "✈": 379,
396
+ "✠": 380,
397
+ "✨": 381,
398
+ "⟙": 382,
399
+ "⸺": 383,
400
+ "。": 384,
401
+ "〈": 385,
402
+ "な": 386,
403
+ "三": 387,
404
+ "次": 388,
405
+ "肿": 389,
406
+ "臃": 390,
407
+ "구": 391,
408
+ "로": 392,
409
+ "요": 393,
410
+ "ﷺ": 394,
411
+ "️": 395,
412
+ "𝑥": 396,
413
+ "𝑦": 397,
414
+ "🌊": 398,
415
+ "🌿": 399,
416
+ "💦": 400,
417
+ "💨": 401
418
+ },
419
+ "phoneme_symbols": {
420
+ "0": "_",
421
+ "1": "<de>",
422
+ "2": "<en_gb>",
423
+ "3": "<en_us>",
424
+ "4": "<es>",
425
+ "5": "<fr>",
426
+ "6": "<it>",
427
+ "7": "<vi>",
428
+ "8": "<end>",
429
+ "9": " ",
430
+ "10": "!",
431
+ "11": ",",
432
+ "12": "-",
433
+ "13": ".",
434
+ "14": "1",
435
+ "15": "2",
436
+ "16": "4",
437
+ "17": "5",
438
+ "18": "6",
439
+ "19": "7",
440
+ "20": ":",
441
+ "21": ";",
442
+ "22": "?",
443
+ "23": "a",
444
+ "24": "b",
445
+ "25": "c",
446
+ "26": "d",
447
+ "27": "e",
448
+ "28": "f",
449
+ "29": "h",
450
+ "30": "i",
451
+ "31": "j",
452
+ "32": "k",
453
+ "33": "l",
454
+ "34": "m",
455
+ "35": "n",
456
+ "36": "o",
457
+ "37": "p",
458
+ "38": "q",
459
+ "39": "r",
460
+ "40": "s",
461
+ "41": "t",
462
+ "42": "u",
463
+ "43": "v",
464
+ "44": "w",
465
+ "45": "x",
466
+ "46": "y",
467
+ "47": "z",
468
+ "48": "«",
469
+ "49": "»",
470
+ "50": "æ",
471
+ "51": "ç",
472
+ "52": "ð",
473
+ "53": "ø",
474
+ "54": "ŋ",
475
+ "55": "œ",
476
+ "56": "ɐ",
477
+ "57": "ɑ",
478
+ "58": "ɒ",
479
+ "59": "ɔ",
480
+ "60": "ɗ",
481
+ "61": "ə",
482
+ "62": "ɚ",
483
+ "63": "ɛ",
484
+ "64": "ɜ",
485
+ "65": "ɟ",
486
+ "66": "ɡ",
487
+ "67": "ɣ",
488
+ "68": "ɨ",
489
+ "69": "ɪ",
490
+ "70": "ɬ",
491
+ "71": "ɯ",
492
+ "72": "ɲ",
493
+ "73": "ɹ",
494
+ "74": "ɾ",
495
+ "75": "ʁ",
496
+ "76": "ʃ",
497
+ "77": "ʊ",
498
+ "78": "ʌ",
499
+ "79": "ʎ",
500
+ "80": "ʐ",
501
+ "81": "ʒ",
502
+ "82": "ʔ",
503
+ "83": "ʝ",
504
+ "84": "ʲ",
505
+ "85": "ˈ",
506
+ "86": "ˌ",
507
+ "87": "ː",
508
+ "88": "̃",
509
+ "89": "̩",
510
+ "90": "̪",
511
+ "91": "β",
512
+ "92": "θ",
513
+ "93": "ᵻ",
514
+ "94": "—",
515
+ "95": "…"
516
+ }
517
+ }
coreml/assets/languages.json ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "id": "en_us",
4
+ "index": 0
5
+ },
6
+ {
7
+ "id": "en_gb",
8
+ "index": 1
9
+ },
10
+ {
11
+ "id": "es",
12
+ "index": 2
13
+ },
14
+ {
15
+ "id": "it",
16
+ "index": 3
17
+ },
18
+ {
19
+ "id": "fr",
20
+ "index": 4
21
+ },
22
+ {
23
+ "id": "de",
24
+ "index": 5
25
+ },
26
+ {
27
+ "id": "vi",
28
+ "index": 6
29
+ }
30
+ ]
coreml/assets/phone_vocab.json ADDED
@@ -0,0 +1,111 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": 1,
3
+ "mode": "scyllasband_ipa_phone_symbols",
4
+ "special_tokens": [
5
+ "<pad>",
6
+ "<unk>",
7
+ "<sil>",
8
+ "<pause_comma>",
9
+ "<pause_semicolon>",
10
+ "<pause_colon>",
11
+ "<pause_dash>",
12
+ "<ellipsis>",
13
+ "<end_stmt>",
14
+ "<end_question>",
15
+ "<end_exclaim>",
16
+ "<ctx_sentence_start>",
17
+ "<ctx_continuation>",
18
+ "<ctx_sentence_end>",
19
+ "<ctx_chunk_continue>"
20
+ ],
21
+ "vocab_size": 87,
22
+ "token_to_id": {
23
+ "<pad>": 0,
24
+ "<unk>": 1,
25
+ "<sil>": 2,
26
+ "<pause_comma>": 3,
27
+ "<pause_semicolon>": 4,
28
+ "<pause_colon>": 5,
29
+ "<pause_dash>": 6,
30
+ "<ellipsis>": 7,
31
+ "<end_stmt>": 8,
32
+ "<end_question>": 9,
33
+ "<end_exclaim>": 10,
34
+ "<ctx_sentence_start>": 11,
35
+ "<ctx_continuation>": 12,
36
+ "<ctx_sentence_end>": 13,
37
+ "<ctx_chunk_continue>": 14,
38
+ "1": 15,
39
+ "2": 16,
40
+ "4": 17,
41
+ "5": 18,
42
+ "6": 19,
43
+ "7": 20,
44
+ "a": 21,
45
+ "b": 22,
46
+ "c": 23,
47
+ "d": 24,
48
+ "e": 25,
49
+ "f": 26,
50
+ "h": 27,
51
+ "i": 28,
52
+ "j": 29,
53
+ "k": 30,
54
+ "l": 31,
55
+ "m": 32,
56
+ "n": 33,
57
+ "o": 34,
58
+ "p": 35,
59
+ "r": 36,
60
+ "s": 37,
61
+ "t": 38,
62
+ "u": 39,
63
+ "v": 40,
64
+ "w": 41,
65
+ "x": 42,
66
+ "y": 43,
67
+ "z": 44,
68
+ "æ": 45,
69
+ "ç": 46,
70
+ "ð": 47,
71
+ "ø": 48,
72
+ "ŋ": 49,
73
+ "œ": 50,
74
+ "ɐ": 51,
75
+ "ɑ": 52,
76
+ "ɒ": 53,
77
+ "ɔ": 54,
78
+ "ɗ": 55,
79
+ "ə": 56,
80
+ "ɚ": 57,
81
+ "ɛ": 58,
82
+ "ɜ": 59,
83
+ "ɟ": 60,
84
+ "ɡ": 61,
85
+ "ɣ": 62,
86
+ "ɪ": 63,
87
+ "ɯ": 64,
88
+ "ɲ": 65,
89
+ "ɹ": 66,
90
+ "ɾ": 67,
91
+ "ʁ": 68,
92
+ "ʃ": 69,
93
+ "ʊ": 70,
94
+ "ʌ": 71,
95
+ "ʎ": 72,
96
+ "ʐ": 73,
97
+ "ʒ": 74,
98
+ "ʔ": 75,
99
+ "ʝ": 76,
100
+ "ʲ": 77,
101
+ "ˈ": 78,
102
+ "ˌ": 79,
103
+ "ː": 80,
104
+ "̃": 81,
105
+ "̩": 82,
106
+ "̪": 83,
107
+ "β": 84,
108
+ "θ": 85,
109
+ "ᵻ": 86
110
+ }
111
+ }
coreml/assets/voices.json ADDED
@@ -0,0 +1,58 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "id": "ariadne",
4
+ "index": 0
5
+ },
6
+ {
7
+ "id": "felix",
8
+ "index": 1
9
+ },
10
+ {
11
+ "id": "gwen",
12
+ "index": 2
13
+ },
14
+ {
15
+ "id": "ink",
16
+ "index": 3
17
+ },
18
+ {
19
+ "id": "max",
20
+ "index": 4
21
+ },
22
+ {
23
+ "id": "orpheus",
24
+ "index": 5
25
+ },
26
+ {
27
+ "id": "rex",
28
+ "index": 6
29
+ },
30
+ {
31
+ "id": "scylla",
32
+ "index": 7
33
+ },
34
+ {
35
+ "id": "stone",
36
+ "index": 8
37
+ },
38
+ {
39
+ "id": "tuesday",
40
+ "index": 9
41
+ },
42
+ {
43
+ "id": "thistle",
44
+ "index": 10
45
+ },
46
+ {
47
+ "id": "kincade",
48
+ "index": 11
49
+ },
50
+ {
51
+ "id": "poppy",
52
+ "index": 12
53
+ },
54
+ {
55
+ "id": "priya",
56
+ "index": 13
57
+ }
58
+ ]
coreml/coreml/duration_predictor.mlmodelc/analytics/coremldata.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ff481a2cd37fa03e16ecda507b033ec1f76b69469819a3259e3574bf9ebe72ab
3
+ size 243
coreml/coreml/duration_predictor.mlmodelc/coremldata.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:06bfdf64690cc0de4e97a0af486c678bf498b4b4224159fd40af1ba8ae296c60
3
+ size 554
coreml/coreml/duration_predictor.mlmodelc/model.mil ADDED
The diff for this file is too large to render. See raw diff
 
coreml/coreml/duration_predictor.mlmodelc/weights/weight.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e6314ef13cc2edaffdc96840a5bf4c7e729630fb22e562f5383be48e3458038
3
+ size 10580928
coreml/coreml/g2p.mlmodelc/analytics/coremldata.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9047c70b85b77b6d386f1e6188323f82daf023854c2ae4c7018f8406c04320fb
3
+ size 243
coreml/coreml/g2p.mlmodelc/coremldata.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5fc8bee0186f33fa14d35a59a58ec3393676d07bdf03802a3836d806662cd03
3
+ size 377