Publish ZeroTTS v1

.gitattributes

@@ -1,35 +1,6 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.data filter=lfs diff=lfs merge=lfs -text
 *.npy filter=lfs diff=lfs merge=lfs -text
 *.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,157 @@
+---
+license: mit
+language:
+  - vi
+library_name: onnx
+pipeline_tag: text-to-speech
+tags:
+  - text-to-speech
+  - tts
+  - vietnamese
+  - onnx
+  - onnxruntime
+  - zero-shot
+  - speech-synthesis
+---
+
+# ZeroTTS
+
+Vietnamese text-to-speech. The whole inference path is **numpy + ONNX Runtime** —
+no PyTorch, no CUDA — so it runs on a laptop CPU or in a browser.
+
+* Code, examples, browser demo: **https://github.com/zeroweight-ai/ZeroTTS**
+* Benchmark dataset: **https://huggingface.co/datasets/zeroweight-ai/ZeroBench-TTS**
+
+```python
+pip install zerotts
+```
+
+```python
+from zerotts import ZeroTTS
+
+tts = ZeroTTS.from_pretrained("zeroweight-ai/ZeroTTS")
+audio = tts.synthesize("Xin chào các bạn, mình là ZeroTTS.", voice="arya")
+tts.save_audio(audio, "out.wav")
+```
+
+Streaming, with first audio in roughly 100 ms:
+
+```python
+for chunk in tts.synthesize_stream("Một đoạn văn bản dài hơn…", voice="arya"):
+    play(chunk)   # (1, n) float32 at 48 kHz
+```
+
+## Benchmarks
+
+[ZeroBench-TTS](https://huggingface.co/datasets/zeroweight-ai/ZeroBench-TTS) —
+137 items, 59 held-out reference voices × 4 subsets, against the two public
+Vietnamese XTTS finetunes. 137/137 scored, 0 empty generations.
+
+WER uses [`vinai/PhoWhisper-large`](https://huggingface.co/vinai/PhoWhisper-large)
+(a Vietnamese-specialized Whisper finetune — scoring Vietnamese TTS with raw
+`whisper-large-v3` measures the ASR's weakness as much as the TTS system's).
+Every model is reported twice: `as-is`, and `+vinorm` which adds
+[`soe-vinorm`](https://pypi.org/project/soe-vinorm/)'s automatic spoken-out
+normalization as an extra accepted reference.
+
+| Model | WER as-is | WER +vinorm | SSIM | UTMOS | Excess silence |
+|---|---|---|---|---|---|
+| **ZeroTTS** | **5.09%** | **4.51%** | 0.936 | **2.95** | **0.029 s** |
+| XTTS-v2-vietnamse | 21.49% | 21.16% | **0.940** | 2.36 | 0.532 s |
+| viXTTS | 25.13% | 24.92% | 0.935 | 2.35 | 0.233 s |
+
+Per subset (WER, as-is → +vinorm):
+
+| Subset | n | ZeroTTS | XTTS-v2-vietnamse | viXTTS |
+|---|---|---|---|---|
+| `vietnamese` | 39 | **0.16% → 0.16%** | 8.18% → 8.18% | 11.25% → 11.25% |
+| `code_switch` | 39 | **5.74% → 5.72%** | 16.16% → 16.15% | 17.66% → 17.64% |
+| `cross_lingual` (foreign voice, vi text) | 20 | **6.05% → 5.98%** | 30.49% → 30.37% | 34.51% → 34.51% |
+| `challenging` (acronyms, dates, %) | 39 | **8.86% → 6.90%** | 35.50% → 34.43% | 41.68% → 40.93% |
+
+Reading these fairly:
+
+* **The `+vinorm` column tests an objection rather than flattering us.** XTTS's
+  stock tokenizer has no Vietnamese number/symbol expansion, so one could argue
+  the `challenging` gap is a missing text frontend. Supplying that normalization
+  on the scoring side moves XTTS 35.50% → 34.43% and viXTTS 41.68% → 40.93% —
+  about a point each. It does not close a 26-point gap.
+* **Voice similarity is a tie, not a win** (0.936 / 0.940 / 0.935, within noise),
+  and on `cross_lingual` ZeroTTS is genuinely behind (0.911 vs ~0.935).
+
+Full tables, ablations and reproduction commands:
+[docs/BENCHMARKS.md](https://github.com/zeroweight-ai/ZeroTTS/blob/main/docs/BENCHMARKS.md).
+
+## Voices, and voice cloning
+
+A voice is a small array of speaker latents, `(1, n_voice_queries, d_model)`,
+shipped as a `.npz` under `voices/`. That array is the entire speaker
+conditioning — no reference transcript, no audio prompt.
+
+> **Voice cloning is not available in this release.** Those latents come from a
+> voice encoder that reads a reference clip, and that encoder is not published.
+> This repository ships ready-to-use voices; it cannot create new ones from
+> audio.
+>
+> To get latents for your own speaker, see **[zeroweight.ai](https://zeroweight.ai)**
+> or get in touch.
+
+Because a voice is just an array, latents obtained that way drop into
+`voices/<name>/voice.npz` and work with no code change.
+
+## Repository layout
+
+```
+config.json                          runtime config
+tokenizer.json                       BPE tokenizer
+null_voice_emb.npy                   learned unconditional voice prefix
+onnx/text_encoder.onnx               text → encoder states (once per utterance)
+onnx/prefix_step.onnx                global transformer step (once per frame)
+onnx/local_frame_decode.onnx         frame decode + sampling (once per frame)
+onnx/codec/                          MOSS-Audio-Tokenizer-Nano decoder (Apache-2.0)
+voices/<name>/voice.npz              speaker latents
+```
+
+fp32, not quantized: ~900 MB total. Two ONNX Runtime calls per audio frame;
+frames are produced at 12.5 Hz and decoded to 48 kHz.
+
+The model architecture, training code, and the ONNX export script are not
+published, and the voice encoder is not included.
+
+## Intended use and limitations
+
+Built for **Vietnamese**. It handles English words embedded in Vietnamese text
+(`code_switch`), but it is not an English TTS system and is not evaluated as one.
+
+Do not use it to impersonate a real person, to generate speech attributed to
+someone without their consent, or to produce audio intended to deceive. The
+shipped voices are for evaluation and demos.
+
+Synthetic speech should be disclosed as synthetic wherever a listener might
+reasonably assume otherwise.
+
+## Credits
+
+Speech codec: **MOSS-Audio-Tokenizer-Nano** by the OpenMOSS team, Apache-2.0.
+Its ONNX **decoder** graphs are redistributed under `onnx/codec/` so ZeroTTS has
+no external runtime dependency; the encoder is not included. See
+`onnx/codec/LICENSE-Apache-2.0.txt`.
+
+```bibtex
+@misc{gong2026mossaudiotokenizerscalingaudiotokenizers,
+  title={MOSS-Audio-Tokenizer: Scaling Audio Tokenizers for Future Audio Foundation Models},
+  author={Yitian Gong and Kuangwei Chen and Zhaoye Fei and Xiaogui Yang and Ke Chen
+          and Yang Wang and Kexin Huang and Mingshu Chen and Ruixiao Li
+          and Qingyuan Cheng and Shimin Li and Xipeng Qiu},
+  year={2026}, eprint={2602.10934}, archivePrefix={arXiv}, primaryClass={cs.SD}
+}
+```
+
+## License
+
+ZeroTTS weights and code: **MIT**.
+Bundled MOSS codec decoder: **Apache-2.0**.
+
+The ZeroBench-TTS *dataset* is CC-BY-NC-4.0 because it redistributes reference
+audio from VIVOS, viVoice, phoaudiobook and Emilia. That license applies to the
+benchmark dataset only — **not** to these weights.

config.json

@@ -0,0 +1,22 @@
+{
+  "text_format": "bpe",
+  "vocab_size": 8192,
+  "num_codebooks": 16,
+  "codebook_size": 1024,
+  "d_model": 768,
+  "n_heads": 12,
+  "n_layers": 9,
+  "n_voice_queries": 10,
+  "special_tokens": {
+    "<pad>": 0,
+    "<bos>": 1,
+    "<eot>": 2,
+    "<soa>": 3,
+    "<slot>": 4,
+    "<eoa>": 5,
+    "<en>": 6,
+    "<vi>": 7
+  },
+  "sample_rate": 48000,
+  "codec_frame_rate": 12.5
+}

null_voice_emb.npy

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec014c14f79e8fc16f4ca6ea557f33fa2cbd7d448b5de5f467aace9ca0321e9f
+size 30848

onnx/codec/LICENSE-Apache-2.0.txt

@@ -0,0 +1,9 @@
+MOSS-Audio-Tokenizer-Nano
+Copyright the OpenMOSS Team.
+Licensed under the Apache License, Version 2.0.
+https://www.apache.org/licenses/LICENSE-2.0
+
+Source: https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX
+        https://github.com/OpenMOSS/MOSS-Audio-Tokenizer
+
+Only the DECODER graphs are redistributed here; the encoder is not included.

onnx/codec/codec_browser_onnx_meta.json

@@ -0,0 +1,564 @@
+{
+ "format_version": 2,
+ "checkpoint_path": "MOSS-Audio-Tokenizer-Nano",
+ "files": {
+  "decode_full": "moss_audio_tokenizer_decode_full.onnx",
+  "decode_step": "moss_audio_tokenizer_decode_step.onnx"
+ },
+ "external_data_files": {
+  "moss_audio_tokenizer_decode_full.onnx": [
+   "moss_audio_tokenizer_decode_shared.data"
+  ],
+  "moss_audio_tokenizer_decode_step.onnx": [
+   "moss_audio_tokenizer_decode_shared.data"
+  ]
+ },
+ "codec_config": {
+  "sample_rate": 48000,
+  "channels": 2,
+  "downsample_rate": 3840,
+  "num_quantizers": 16
+ },
+ "onnx": {
+  "opset": 17,
+  "decode_input_names": [
+   "audio_codes",
+   "audio_code_lengths"
+  ],
+  "decode_output_names": [
+   "audio",
+   "audio_lengths"
+  ],
+  "decode_step_input_names": [
+   "audio_codes",
+   "audio_code_lengths",
+   "transformer_offset_0",
+   "transformer_offset_1",
+   "transformer_offset_2",
+   "transformer_offset_3",
+   "attn_offset_0",
+   "attn_cached_keys_0",
+   "attn_cached_values_0",
+   "attn_cached_positions_0",
+   "attn_offset_1",
+   "attn_cached_keys_1",
+   "attn_cached_values_1",
+   "attn_cached_positions_1",
+   "attn_offset_2",
+   "attn_cached_keys_2",
+   "attn_cached_values_2",
+   "attn_cached_positions_2",
+   "attn_offset_3",
+   "attn_cached_keys_3",
+   "attn_cached_values_3",
+   "attn_cached_positions_3",
+   "attn_offset_4",
+   "attn_cached_keys_4",
+   "attn_cached_values_4",
+   "attn_cached_positions_4",
+   "attn_offset_5",
+   "attn_cached_keys_5",
+   "attn_cached_values_5",
+   "attn_cached_positions_5",
+   "attn_offset_6",
+   "attn_cached_keys_6",
+   "attn_cached_values_6",
+   "attn_cached_positions_6",
+   "attn_offset_7",
+   "attn_cached_keys_7",
+   "attn_cached_values_7",
+   "attn_cached_positions_7",
+   "attn_offset_8",
+   "attn_cached_keys_8",
+   "attn_cached_values_8",
+   "attn_cached_positions_8",
+   "attn_offset_9",
+   "attn_cached_keys_9",
+   "attn_cached_values_9",
+   "attn_cached_positions_9",
+   "attn_offset_10",
+   "attn_cached_keys_10",
+   "attn_cached_values_10",
+   "attn_cached_positions_10",
+   "attn_offset_11",
+   "attn_cached_keys_11",
+   "attn_cached_values_11",
+   "attn_cached_positions_11"
+  ],
+  "decode_step_output_names": [
+   "audio",
+   "audio_lengths",
+   "transformer_offset_out_0",
+   "transformer_offset_out_1",
+   "transformer_offset_out_2",
+   "transformer_offset_out_3",
+   "attn_offset_out_0",
+   "attn_cached_keys_out_0",
+   "attn_cached_values_out_0",
+   "attn_cached_positions_out_0",
+   "attn_offset_out_1",
+   "attn_cached_keys_out_1",
+   "attn_cached_values_out_1",
+   "attn_cached_positions_out_1",
+   "attn_offset_out_2",
+   "attn_cached_keys_out_2",
+   "attn_cached_values_out_2",
+   "attn_cached_positions_out_2",
+   "attn_offset_out_3",
+   "attn_cached_keys_out_3",
+   "attn_cached_values_out_3",
+   "attn_cached_positions_out_3",
+   "attn_offset_out_4",
+   "attn_cached_keys_out_4",
+   "attn_cached_values_out_4",
+   "attn_cached_positions_out_4",
+   "attn_offset_out_5",
+   "attn_cached_keys_out_5",
+   "attn_cached_values_out_5",
+   "attn_cached_positions_out_5",
+   "attn_offset_out_6",
+   "attn_cached_keys_out_6",
+   "attn_cached_values_out_6",
+   "attn_cached_positions_out_6",
+   "attn_offset_out_7",
+   "attn_cached_keys_out_7",
+   "attn_cached_values_out_7",
+   "attn_cached_positions_out_7",
+   "attn_offset_out_8",
+   "attn_cached_keys_out_8",
+   "attn_cached_values_out_8",
+   "attn_cached_positions_out_8",
+   "attn_offset_out_9",
+   "attn_cached_keys_out_9",
+   "attn_cached_values_out_9",
+   "attn_cached_positions_out_9",
+   "attn_offset_out_10",
+   "attn_cached_keys_out_10",
+   "attn_cached_values_out_10",
+   "attn_cached_positions_out_10",
+   "attn_offset_out_11",
+   "attn_cached_keys_out_11",
+   "attn_cached_values_out_11",
+   "attn_cached_positions_out_11"
+  ]
+ },
+ "streaming_decode": {
+  "batch_size": 1,
+  "transformer_offsets": [
+   {
+    "index": 0,
+    "decoder_index": 1,
+    "input_name": "transformer_offset_0",
+    "output_name": "transformer_offset_out_0",
+    "shape": [
+     1
+    ],
+    "dtype": "int32"
+   },
+   {
+    "index": 1,
+    "decoder_index": 3,
+    "input_name": "transformer_offset_1",
+    "output_name": "transformer_offset_out_1",
+    "shape": [
+     1
+    ],
+    "dtype": "int32"
+   },
+   {
+    "index": 2,
+    "decoder_index": 5,
+    "input_name": "transformer_offset_2",
+    "output_name": "transformer_offset_out_2",
+    "shape": [
+     1
+    ],
+    "dtype": "int32"
+   },
+   {
+    "index": 3,
+    "decoder_index": 7,
+    "input_name": "transformer_offset_3",
+    "output_name": "transformer_offset_out_3",
+    "shape": [
+     1
+    ],
+    "dtype": "int32"
+   }
+  ],
+  "attention_caches": [
+   {
+    "index": 0,
+    "decoder_index": 1,
+    "layer_index": 0,
+    "context": 500,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_0",
+    "offset_output_name": "attn_offset_out_0",
+    "cached_keys_input_name": "attn_cached_keys_0",
+    "cached_keys_output_name": "attn_cached_keys_out_0",
+    "cached_values_input_name": "attn_cached_values_0",
+    "cached_values_output_name": "attn_cached_values_out_0",
+    "cached_positions_input_name": "attn_cached_positions_0",
+    "cached_positions_output_name": "attn_cached_positions_out_0",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     500,
+     64
+    ],
+    "positions_shape": [
+     1,
+     500
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 1,
+    "decoder_index": 1,
+    "layer_index": 1,
+    "context": 500,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_1",
+    "offset_output_name": "attn_offset_out_1",
+    "cached_keys_input_name": "attn_cached_keys_1",
+    "cached_keys_output_name": "attn_cached_keys_out_1",
+    "cached_values_input_name": "attn_cached_values_1",
+    "cached_values_output_name": "attn_cached_values_out_1",
+    "cached_positions_input_name": "attn_cached_positions_1",
+    "cached_positions_output_name": "attn_cached_positions_out_1",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     500,
+     64
+    ],
+    "positions_shape": [
+     1,
+     500
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 2,
+    "decoder_index": 1,
+    "layer_index": 2,
+    "context": 500,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_2",
+    "offset_output_name": "attn_offset_out_2",
+    "cached_keys_input_name": "attn_cached_keys_2",
+    "cached_keys_output_name": "attn_cached_keys_out_2",
+    "cached_values_input_name": "attn_cached_values_2",
+    "cached_values_output_name": "attn_cached_values_out_2",
+    "cached_positions_input_name": "attn_cached_positions_2",
+    "cached_positions_output_name": "attn_cached_positions_out_2",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     500,
+     64
+    ],
+    "positions_shape": [
+     1,
+     500
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 3,
+    "decoder_index": 1,
+    "layer_index": 3,
+    "context": 500,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_3",
+    "offset_output_name": "attn_offset_out_3",
+    "cached_keys_input_name": "attn_cached_keys_3",
+    "cached_keys_output_name": "attn_cached_keys_out_3",
+    "cached_values_input_name": "attn_cached_values_3",
+    "cached_values_output_name": "attn_cached_values_out_3",
+    "cached_positions_input_name": "attn_cached_positions_3",
+    "cached_positions_output_name": "attn_cached_positions_out_3",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     500,
+     64
+    ],
+    "positions_shape": [
+     1,
+     500
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 4,
+    "decoder_index": 3,
+    "layer_index": 0,
+    "context": 800,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_4",
+    "offset_output_name": "attn_offset_out_4",
+    "cached_keys_input_name": "attn_cached_keys_4",
+    "cached_keys_output_name": "attn_cached_keys_out_4",
+    "cached_values_input_name": "attn_cached_values_4",
+    "cached_values_output_name": "attn_cached_values_out_4",
+    "cached_positions_input_name": "attn_cached_positions_4",
+    "cached_positions_output_name": "attn_cached_positions_out_4",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     800,
+     64
+    ],
+    "positions_shape": [
+     1,
+     800
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 5,
+    "decoder_index": 3,
+    "layer_index": 1,
+    "context": 800,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_5",
+    "offset_output_name": "attn_offset_out_5",
+    "cached_keys_input_name": "attn_cached_keys_5",
+    "cached_keys_output_name": "attn_cached_keys_out_5",
+    "cached_values_input_name": "attn_cached_values_5",
+    "cached_values_output_name": "attn_cached_values_out_5",
+    "cached_positions_input_name": "attn_cached_positions_5",
+    "cached_positions_output_name": "attn_cached_positions_out_5",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     800,
+     64
+    ],
+    "positions_shape": [
+     1,
+     800
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 6,
+    "decoder_index": 5,
+    "layer_index": 0,
+    "context": 1200,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_6",
+    "offset_output_name": "attn_offset_out_6",
+    "cached_keys_input_name": "attn_cached_keys_6",
+    "cached_keys_output_name": "attn_cached_keys_out_6",
+    "cached_values_input_name": "attn_cached_values_6",
+    "cached_values_output_name": "attn_cached_values_out_6",
+    "cached_positions_input_name": "attn_cached_positions_6",
+    "cached_positions_output_name": "attn_cached_positions_out_6",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     1200,
+     64
+    ],
+    "positions_shape": [
+     1,
+     1200
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 7,
+    "decoder_index": 5,
+    "layer_index": 1,
+    "context": 1200,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_7",
+    "offset_output_name": "attn_offset_out_7",
+    "cached_keys_input_name": "attn_cached_keys_7",
+    "cached_keys_output_name": "attn_cached_keys_out_7",
+    "cached_values_input_name": "attn_cached_values_7",
+    "cached_values_output_name": "attn_cached_values_out_7",
+    "cached_positions_input_name": "attn_cached_positions_7",
+    "cached_positions_output_name": "attn_cached_positions_out_7",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     1200,
+     64
+    ],
+    "positions_shape": [
+     1,
+     1200
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 8,
+    "decoder_index": 7,
+    "layer_index": 0,
+    "context": 1600,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_8",
+    "offset_output_name": "attn_offset_out_8",
+    "cached_keys_input_name": "attn_cached_keys_8",
+    "cached_keys_output_name": "attn_cached_keys_out_8",
+    "cached_values_input_name": "attn_cached_values_8",
+    "cached_values_output_name": "attn_cached_values_out_8",
+    "cached_positions_input_name": "attn_cached_positions_8",
+    "cached_positions_output_name": "attn_cached_positions_out_8",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     1600,
+     64
+    ],
+    "positions_shape": [
+     1,
+     1600
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 9,
+    "decoder_index": 7,
+    "layer_index": 1,
+    "context": 1600,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_9",
+    "offset_output_name": "attn_offset_out_9",
+    "cached_keys_input_name": "attn_cached_keys_9",
+    "cached_keys_output_name": "attn_cached_keys_out_9",
+    "cached_values_input_name": "attn_cached_values_9",
+    "cached_values_output_name": "attn_cached_values_out_9",
+    "cached_positions_input_name": "attn_cached_positions_9",
+    "cached_positions_output_name": "attn_cached_positions_out_9",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     1600,
+     64
+    ],
+    "positions_shape": [
+     1,
+     1600
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 10,
+    "decoder_index": 7,
+    "layer_index": 2,
+    "context": 1600,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_10",
+    "offset_output_name": "attn_offset_out_10",
+    "cached_keys_input_name": "attn_cached_keys_10",
+    "cached_keys_output_name": "attn_cached_keys_out_10",
+    "cached_values_input_name": "attn_cached_values_10",
+    "cached_values_output_name": "attn_cached_values_out_10",
+    "cached_positions_input_name": "attn_cached_positions_10",
+    "cached_positions_output_name": "attn_cached_positions_out_10",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     1600,
+     64
+    ],
+    "positions_shape": [
+     1,
+     1600
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   },
+   {
+    "index": 11,
+    "decoder_index": 7,
+    "layer_index": 3,
+    "context": 1600,
+    "num_heads": 4,
+    "head_dim": 64,
+    "offset_input_name": "attn_offset_11",
+    "offset_output_name": "attn_offset_out_11",
+    "cached_keys_input_name": "attn_cached_keys_11",
+    "cached_keys_output_name": "attn_cached_keys_out_11",
+    "cached_values_input_name": "attn_cached_values_11",
+    "cached_values_output_name": "attn_cached_values_out_11",
+    "cached_positions_input_name": "attn_cached_positions_11",
+    "cached_positions_output_name": "attn_cached_positions_out_11",
+    "offset_shape": [
+     1
+    ],
+    "cache_shape": [
+     1,
+     4,
+     1600,
+     64
+    ],
+    "positions_shape": [
+     1,
+     1600
+    ],
+    "cache_dtype": "float32",
+    "positions_dtype": "int32"
+   }
+  ]
+ }
+}

onnx/codec/moss_audio_tokenizer_decode_full.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0fbbafe3fd4afa2a019af5c5ced204af6e2d1db044fa40f021525d2aee95b4ac
+size 681902

onnx/codec/moss_audio_tokenizer_decode_shared.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e69d52e0f4e84ca27850557ee54face46632d3a5a16c89bd246c7c408466dcad
+size 44198912

onnx/codec/moss_audio_tokenizer_decode_step.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9527c86a29e1837edec1f74db57d5eeaadb3a715af3382703566460afed25855
+size 351400

onnx/local_frame_decode.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c4540ef4e69dcf604dc6f60f6f99f119189694ff5f31113def20406e274f025
+size 186743011

onnx/prefix_step.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7590e502e32e31cc8afd375a6113c974fd7524efe9d9b9b5b5ed63186607f235
+size 391035953

onnx/text_encoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b14e0e3df7431c964bcecbb8aa74cf93f7f6daf70cec04cf157ac539a7bf485c
+size 280483795

voices/arya/meta.json

@@ -0,0 +1,7 @@
+{
+  "name": "arya",
+  "language": "vi",
+  "description": "Giọng nữ, rõ ràng và trung tính",
+  "gender": "female",
+  "n_voice_queries": 10
+}

voices/arya/preview.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89b762d566dc7cc734059114d955182f8f30341f6ccbc2c240eaa4ef98f31e2a
+size 459820

voices/arya/voice.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ad725aed27afff9487c5a00c715027df481c93922028b698e1d61c59e134ff36
+size 30720

voices/arya/voice.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dee86069210c4b59e70d548c96830fe61d809c79ce321755f5f21e08b615473e
+size 31262

voices/index.json

@@ -0,0 +1,26 @@
+{
+  "voices": [
+    {
+      "name": "arya",
+      "language": "vi",
+      "description": "Giọng nữ, rõ ràng và trung tính",
+      "n_voice_queries": 10,
+      "shape": [
+        1,
+        10,
+        768
+      ]
+    },
+    {
+      "name": "viVoice1",
+      "language": "vi",
+      "description": "Giọng đọc tự nhiên, tốc độ vừa phải",
+      "n_voice_queries": 10,
+      "shape": [
+        1,
+        10,
+        768
+      ]
+    }
+  ]
+}

voices/viVoice1/meta.json

@@ -0,0 +1,7 @@
+{
+  "name": "viVoice1",
+  "language": "vi",
+  "description": "Giọng đọc tự nhiên, tốc độ vừa phải",
+  "gender": "",
+  "n_voice_queries": 10
+}

voices/viVoice1/preview.wav

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:129face8e1bb175b43bfa6baabae3f3bbf61d5443199b09f01bb1770317a7992
+size 576044

voices/viVoice1/voice.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:552f9fd7cc6cc5b2161a9b5248c8dd6745bf227b537bdbab4b5bfe14f2ea0df1
+size 30720

voices/viVoice1/voice.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30051535e078df5d6a7547c4b90a175c5a42aea4142c7122efc03a5e2c145f1d
+size 31262