roxi-tts-v2 ONNX/CPU build (tuned sampling; IIT-M attribution)

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* 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
+moss_tts_global_shared.data filter=lfs diff=lfs merge=lfs -text
+moss_tts_local_shared.data filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,71 @@
+---
+license: apache-2.0
+language:
+- en
+pipeline_tag: text-to-speech
+base_model: IOTEverythin/roxi-tts-v2
+tags:
+- text-to-speech
+- tts
+- onnx
+- cpu
+- moss-tts-nano
+- indian-english
+---
+
+# Roxi-TTS v2 — ONNX / CPU build
+
+**ONNX export of [`IOTEverythin/roxi-tts-v2`](https://huggingface.co/IOTEverythin/roxi-tts-v2)**
+(MOSS-TTS-Nano LoRA, Indian-English) for **CPU / on-device** inference via `onnxruntime` —
+no PyTorch required at inference. 48 kHz output.
+
+> Use this **ONNX** build for fast CPU serving. For **maximum voice fidelity**, use the PyTorch
+> build [`IOTEverythin/roxi-tts-v2`](https://huggingface.co/IOTEverythin/roxi-tts-v2) (see quality note).
+
+## Contents
+Browser/CPU ONNX bundle exported with MOSS-TTS-Nano's official exporter:
+`moss_tts_prefill.onnx`, `moss_tts_decode_step.onnx`, `moss_tts_local_*` graphs,
+shared external data (`*_shared.data`), `tokenizer.model`, `browser_poc_manifest.json`,
+`tts_browser_onnx_meta.json`. The audio **codec** is the official Apache-2.0
+[`OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX`](https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX)
+(not re-hosted here).
+
+## Usage (onnxruntime, CPU)
+Uses the runtime from the [MOSS-TTS-Nano repo](https://github.com/OpenMOSS/MOSS-TTS-Nano)
+(`infer_onnx.py`). Lay the assets out as `<dir>/MOSS-TTS-Nano-100M-ONNX` + `<dir>/MOSS-Audio-Tokenizer-Nano-ONNX`:
+```bash
+pip install "onnxruntime>=1.20" soundfile numpy
+git clone https://github.com/OpenMOSS/MOSS-TTS-Nano && cd MOSS-TTS-Nano
+huggingface-cli download IOTEverythin/roxi-tts-v2-onnx --local-dir assets/MOSS-TTS-Nano-100M-ONNX
+huggingface-cli download OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX --local-dir assets/MOSS-Audio-Tokenizer-Nano-ONNX
+
+python infer_onnx.py --model-dir assets \
+  --text "Welcome. Your appointment is confirmed for Monday at ten thirty in the morning." \
+  --output-audio-path out.wav --execution-provider cpu --cpu-threads 8 \
+  --sample-mode full --audio-temperature 0.8 --audio-top-p 0.9 --audio-repetition-penalty 1.1 \
+  --disable-wetext-processing
+```
+**Recommended sampling (tuned for voice fidelity):** `--sample-mode full --audio-temperature 0.8
+--audio-top-p 0.9 --audio-repetition-penalty 1.1`. (Greedy collapses the voice; the default
+`fixed` mode is lower fidelity.)
+
+## Quality (measured, CPU, vs the PyTorch build)
+| Build | Speaker similarity to target ↑ | Intelligibility WER ↓ |
+|---|---|---|
+| roxi-tts-v2 (PyTorch) | **0.96** | 0.26 |
+| **roxi-tts-v2-onnx (this, tuned)** | **0.73** | **0.25** |
+| base MOSS (reference) | 0.52 | 0.26 |
+
+- **Intelligibility is on par** with PyTorch (WER ~0.25).
+- **Voice identity is somewhat weaker than PyTorch** (0.73 vs 0.96) — the ONNX streaming/fixed-frame
+  decode + fp32 approximate the PyTorch generation path. It is still clearly the fine-tuned
+  Indian-English voice (well above base 0.52). Use the PyTorch build if you need full fidelity.
+- **Speed:** ONNX is the faster CPU path (independent reports ~RTF 0.23–0.28 @ 8 threads, ~2× the
+  PyTorch CPU path) once loaded in a persistent server.
+
+## Attribution & license
+- Apache-2.0 (matching base). Built on **MOSS-TTS-Nano** (Apache-2.0) + audio tokenizer (Apache-2.0).
+- Training data: IIT-Madras **Indic TTS** (English) via `SPRINGLab/IndicTTS-English`. Required notice:
+  *"COPYRIGHT 2016 TTS Consortium, TDIL, Meity — Hema A. Murthy & S. Umesh — IIT Madras. ALL RIGHTS RESERVED."*
+- **Responsible use:** derived from a real dataset speaker — do not use to impersonate real people,
+  for fraud, or deception; disclose AI-generated audio where required.

moss_tts_decode_step.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:698cbc2fc1c2feca16e5895614ed52bbb32ded10f236c076f477b2e69abf32d8
+size 291483

moss_tts_global_shared.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5044370e12c754f82560690db3f270e019ab29ce4f53c75c5ad4d74c84b63fc
+size 440813568

moss_tts_local_cached_step.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa9035fefc1c138a951a8bcfc0374fb03a25f1ece67f7f7f53bce349b84a1dd5
+size 53685

moss_tts_local_decoder.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:51aa754301b38550a5f9adda0ad93bd3dc95819afb511e6dcabf4a90b345a454
+size 49231

moss_tts_local_fixed_sampled_frame.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40cdb00efc171c450cf91468e01429caa41b0252222cd308e978f58fe354afa8
+size 471262

moss_tts_local_shared.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fd95020ca151295fa2ee4eac114785ee2c6ee055a179a7ea5b3d1c6b5e306691
+size 229678080

moss_tts_prefill.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d56126dcd0574c2f15d98fc6b35eda68d0386b5bd9c5e38e28548d6f2ea8f3db
+size 283305

tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c353ee1479b536bf414c1b247f5542b6607fb8ae91320e5af1781fee200fddff
+size 470897

tts_browser_onnx_meta.json

@@ -0,0 +1,186 @@
+{
+  "format_version": 1,
+  "checkpoint_path": "merged-ep5",
+  "files": {
+    "prefill": "moss_tts_prefill.onnx",
+    "decode_step": "moss_tts_decode_step.onnx",
+    "local_decoder": "moss_tts_local_decoder.onnx",
+    "local_cached_step": "moss_tts_local_cached_step.onnx",
+    "local_fixed_sampled_frame": "moss_tts_local_fixed_sampled_frame.onnx"
+  },
+  "external_data_files": {
+    "moss_tts_prefill.onnx": [
+      "moss_tts_global_shared.data"
+    ],
+    "moss_tts_decode_step.onnx": [
+      "moss_tts_global_shared.data"
+    ],
+    "moss_tts_local_decoder.onnx": [
+      "moss_tts_local_shared.data"
+    ],
+    "moss_tts_local_cached_step.onnx": [
+      "moss_tts_local_shared.data"
+    ],
+    "moss_tts_local_fixed_sampled_frame.onnx": [
+      "moss_tts_local_shared.data"
+    ]
+  },
+  "model_config": {
+    "n_vq": 16,
+    "row_width": 17,
+    "hidden_size": 768,
+    "global_layers": 12,
+    "global_heads": 12,
+    "head_dim": 64,
+    "local_layers": 1,
+    "local_heads": 12,
+    "local_head_dim": 64,
+    "vocab_size": 16384,
+    "audio_codebook_sizes": [
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024,
+      1024
+    ],
+    "audio_pad_token_id": 1024,
+    "pad_token_id": 3,
+    "im_start_token_id": 4,
+    "im_end_token_id": 5,
+    "audio_start_token_id": 6,
+    "audio_end_token_id": 7,
+    "audio_user_slot_token_id": 8,
+    "audio_assistant_slot_token_id": 9
+  },
+  "onnx": {
+    "opset": 17,
+    "prefill_output_names": [
+      "global_hidden",
+      "present_key_0",
+      "present_value_0",
+      "present_key_1",
+      "present_value_1",
+      "present_key_2",
+      "present_value_2",
+      "present_key_3",
+      "present_value_3",
+      "present_key_4",
+      "present_value_4",
+      "present_key_5",
+      "present_value_5",
+      "present_key_6",
+      "present_value_6",
+      "present_key_7",
+      "present_value_7",
+      "present_key_8",
+      "present_value_8",
+      "present_key_9",
+      "present_value_9",
+      "present_key_10",
+      "present_value_10",
+      "present_key_11",
+      "present_value_11"
+    ],
+    "decode_input_names": [
+      "input_ids",
+      "past_valid_lengths",
+      "past_key_0",
+      "past_value_0",
+      "past_key_1",
+      "past_value_1",
+      "past_key_2",
+      "past_value_2",
+      "past_key_3",
+      "past_value_3",
+      "past_key_4",
+      "past_value_4",
+      "past_key_5",
+      "past_value_5",
+      "past_key_6",
+      "past_value_6",
+      "past_key_7",
+      "past_value_7",
+      "past_key_8",
+      "past_value_8",
+      "past_key_9",
+      "past_value_9",
+      "past_key_10",
+      "past_value_10",
+      "past_key_11",
+      "past_value_11"
+    ],
+    "decode_output_names": [
+      "global_hidden",
+      "present_key_0",
+      "present_value_0",
+      "present_key_1",
+      "present_value_1",
+      "present_key_2",
+      "present_value_2",
+      "present_key_3",
+      "present_value_3",
+      "present_key_4",
+      "present_value_4",
+      "present_key_5",
+      "present_value_5",
+      "present_key_6",
+      "present_value_6",
+      "present_key_7",
+      "present_value_7",
+      "present_key_8",
+      "present_value_8",
+      "present_key_9",
+      "present_value_9",
+      "present_key_10",
+      "present_value_10",
+      "present_key_11",
+      "present_value_11"
+    ],
+    "local_cached_input_names": [
+      "global_hidden",
+      "text_token_id",
+      "audio_token_id",
+      "channel_index",
+      "step_type",
+      "past_valid_lengths",
+      "local_past_key_0",
+      "local_past_value_0"
+    ],
+    "local_cached_output_names": [
+      "text_logits",
+      "audio_logits",
+      "local_present_key_0",
+      "local_present_value_0"
+    ],
+    "local_fixed_sampled_frame_input_names": [
+      "global_hidden",
+      "repetition_seen_mask",
+      "assistant_random_u",
+      "audio_random_u"
+    ],
+    "local_fixed_sampled_frame_output_names": [
+      "should_continue",
+      "frame_token_ids"
+    ],
+    "fixed_sampled_frame_constants": {
+      "text_temperature": 1.0,
+      "text_top_p": 1.0,
+      "text_top_k": 50,
+      "audio_temperature": 0.8,
+      "audio_top_p": 0.95,
+      "audio_top_k": 25,
+      "audio_repetition_penalty": 1.2
+    }
+  }
+}