Loading
Du hast einen ComfyUI-Workflow? Erstelle direkt im Browser eine verständliche Workflow-Dokumentation.Workflow verständlich machen.Jetzt ausprobieren ↗
A self-contained, fixed-revision Core ML build of the Kokoro-82M 7-stage Apple-Neural-Engine pipeline packaged with all 54 standard Kokoro voices, so a sandboxed Swift / Core ML app can offer every voice on the ANE...
Modellquelle
Quellenauszug
A self-contained, fixed-revision Core ML build of the Kokoro-82M 7-stage Apple-Neural-Engine pipeline packaged with all 54 standard Kokoro voices, so a sandboxed Swift / Core ML app can offer every voice on the ANE...
Quellen
1 QuelleVerifiziert 13. Sept.
Modellartefakte
64 ArtefakteANE/KokoroAlbert.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 14a61873d875…5ccd · Hugging Face
HerunterladenANE/KokoroAlbert.mlmodelc/coremldata.bin
bin · 433 B · SHA-256 70b6d2f84292…63a6 · Hugging Face
HerunterladenQuellenauszüge
2 AuszügeANE/KokoroAlbert.mlmodelc/weights/weight.bin
bin · 5,45 MB · SHA-256 36089a39359b…fabc · Hugging Face
HerunterladenANE/KokoroAlignment.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 f6074d1039a9…4f3f · Hugging Face
HerunterladenANE/KokoroAlignment.mlmodelc/coremldata.bin
bin · 484 B · SHA-256 9a0fb4a536f6…a4ea · Hugging Face
HerunterladenANE/KokoroAlignment.mlmodelc/weights/weight.bin
bin · 4,03 KB · SHA-256 2e7d69128b59…5f6b · Hugging Face
HerunterladenANE/KokoroNoise.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 de1b584b3959…b133 · Hugging Face
HerunterladenANE/KokoroNoise.mlmodelc/coremldata.bin
bin · 440 B · SHA-256 6e675b07b45f…283f · Hugging Face
HerunterladenANE/KokoroNoise.mlmodelc/weights/weight.bin
bin · 4,37 MB · SHA-256 5da4ccb85b78…a0d0 · Hugging Face
HerunterladenANE/KokoroPostAlbert.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 6640044c8755…6c43 · Hugging Face
HerunterladenANE/KokoroPostAlbert.mlmodelc/coremldata.bin
bin · 556 B · SHA-256 86de3ab0c1e8…ba3b · Hugging Face
HerunterladenANE/KokoroPostAlbert.mlmodelc/weights/weight.bin
bin · 13,2 MB · SHA-256 e4f300a23cc2…91c8 · Hugging Face
HerunterladenANE/KokoroProsody.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 7708aab83dea…c6b8 · Hugging Face
HerunterladenANE/KokoroProsody.mlmodelc/coremldata.bin
bin · 421 B · SHA-256 d65d87f246a6…a33e · Hugging Face
HerunterladenANE/KokoroProsody.mlmodelc/weights/weight.bin
bin · 8,06 MB · SHA-256 d3c2670eb0c5…444c · Hugging Face
HerunterladenANE/KokoroTail.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 7dd3d6b8cfbc…77db · Hugging Face
HerunterladenANE/KokoroTail.mlmodelc/coremldata.bin
bin · 392 B · SHA-256 f28f17e4217d…f5d3 · Hugging Face
HerunterladenANE/KokoroTail.mlmodelc/weights/weight.bin
bin · 79,2 KB · SHA-256 2d4877b5d272…ecb2 · Hugging Face
HerunterladenANE/KokoroVocoder.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 8c7c1a25a46a…e74b · Hugging Face
HerunterladenANE/KokoroVocoder.mlmodelc/coremldata.bin
bin · 626 B · SHA-256 e73aaf146c75…6c00 · Hugging Face
HerunterladenANE/KokoroVocoder.mlmodelc/weights/weight.bin
bin · 46,6 MB · SHA-256 6d1f96eb5021…2ff1 · Hugging Face
HerunterladenG2PDecoder.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 dbf1767747fd…9827 · Hugging Face
HerunterladenG2PDecoder.mlmodelc/coremldata.bin
bin · 545 B · SHA-256 607e960f19b4…acd9 · Hugging Face
HerunterladenG2PDecoder.mlmodelc/weights/weight.bin
bin · 809 KB · SHA-256 cbaeb4e74335…8214 · Hugging Face
HerunterladenG2PEncoder.mlmodelc/analytics/coremldata.bin
bin · 243 B · SHA-256 cf7fbd7e7a65…6a99 · Hugging Face
HerunterladenG2PEncoder.mlmodelc/coremldata.bin
bin · 398 B · SHA-256 0f14d46ca9fd…0b11 · Hugging Face
HerunterladenG2PEncoder.mlmodelc/weights/weight.bin
bin · 678 KB · SHA-256 6926bcd2827d…14ec · Hugging Face
Herunterladen--- license: apache-2.0 language: - en - es - fr - hi - it - ja - pt - zh pipeline_tag: text-to-speech base_model: - hexgrad/Kokoro-82M - FluidInference/kokoro-82m-coreml tags: - coreml - kokoro - tts - ane - apple-neural-engine - on-device --- # Kokoro 82M — Core ML ANE chain + all 54 voices A self-contained, fixed-revision Core ML build of the Kokoro-82M 7-stage Apple-Neural-Engine pipeline **packaged with all 54 standard Kokoro voices**, so a sandboxed Swift / Core ML app can offer every voice on the ANE path from a single hash-pinnable download. It mirrors the English ANE chain from [`FluidInference/kokoro-82m-coreml`](https://huggingface.co/FluidInference/kokoro-82m-coreml) and adds the voice packs that repo did not ship in ANE form. This was built for **LocalMacTalk** — a native-Swift, fully sandboxed macOS app for evaluating on-device multimodal models — whose speech-synthesis layer lets the user choose where Kokoro runs: GPU (MLX), the Neural Engine, or the CPU. Moving TTS onto the ANE keeps it off the same Metal device as the language model, which matters when a 12B LLM is decoding and synthesizing speech at the same time. (You don't need LocalMacTalk to use these files — they're plain Core ML models and Kokoro voice packs.) ## Why this repo exists The excellent upstream `FluidInference/kokoro-82m-coreml` ships the 7-stage ANE acoustic chain and the Core ML G2P models, but under `ANE/` it includes only a **single English voice** (`af_heart.bin`). The full standard voice set (`af_*`, `am_*`, `bf_*`, `bm_*`, and the multilingual voices) lives in [`hexgrad/Kokoro-82M`](https://huggingface.co/hexgrad/Kokoro-82M) only as PyTorch `.pt` tensors — not in the flat `[510, 256]` float32 layout the ANE chain consumes. So an app that wants user-selectable voices on the Core ML / ANE path had nothing to download. This repo closes that gap and pins everything to **one fixed commit** (the upstream tracks a moving `main`), so a downstream can hash-verify every artifact. ## What's inside - **7-stage ANE acoustic chain** (`ANE/Kokoro{Albert,PostAlbert,Alignment,Prosody,Noise,Vocoder,Tail}.mlmodelc`) — mirrored from the upstream `ANE/` (English variant of the laishere/kokoro chain). - **Core ML G2P** (`G2PEncoder.mlmodelc`, `G2PDecoder.mlmodelc`, `g2p_vocab.json`) and the phoneme table `ANE/vocab.json` — mirrored from the upstream repo root...
Source context: 0 downloads · 0 likes · Pipeline text-to-speech · Repo txgsync/kokoro-82m-coreml-ane
