Add model card (M16.9)

README.md

@@ -0,0 +1,140 @@
+---
+license: apache-2.0
+language:
+- en
+- zh
+- ja
+library_name: onnx
+tags:
+- text-to-speech
+- tts
+- onnx
+- quantized
+- fp16
+- mobile
+- on-device
+- audio-tokenizer
+- moss
+base_model: OpenMOSS-Team/MOSS-TTS-Nano-100M-ONNX
+pipeline_tag: text-to-speech
+---
+
+# MOSS-TTS-Nano-100M — FP16 (shared external data)
+
+Half-precision-quantized version of
+[OpenMOSS-Team/MOSS-TTS-Nano-100M-ONNX](https://huggingface.co/OpenMOSS-Team/MOSS-TTS-Nano-100M-ONNX),
+re-exported with shared external weight data so **prefill** and **decode_step**
+mmap the same 220 MB blob instead of duplicating weights. Tuned for ARM CPUs
+with native fp16 SIMD (Snapdragon 8 Gen 2/3, Apple A14+, etc).
+
+**Audio codec is unchanged from upstream.** Use this repo for the language
+model graphs and pull the codec from
+[OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX](https://huggingface.co/OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX).
+
+## What's here
+
+| File | Size | Purpose |
+|---|---:|---|
+| `moss_tts_prefill.onnx` | 476 KB | global LM, full-context prefill |
+| `moss_tts_decode_step.onnx` | 503 KB | global LM, autoregressive single step (KV cache) |
+| `moss_tts_local_fixed_sampled_frame.onnx` | 819 KB | local LM with baked-in sampling |
+| `moss_tts_global_shared_fp16.data` | **220 MB** | fp16 weights shared by prefill + decode_step |
+| `moss_tts_local_fixed_sampled_frame_fp16.data` | 113 MB | fp16 weights for the local sampler |
+| `tokenizer.model` | 471 KB | SentencePiece (unchanged from upstream) |
+| `browser_poc_manifest.json` | 503 KB | prompt templates + 18 builtin voices |
+| `tts_browser_onnx_meta.json` | 4 KB | I/O metadata |
+| **Total** | **~336 MB** | (vs 640 MB upstream fp32 LM, 196 MB int8) |
+
+## Variant comparison (this repo's role)
+
+| Variant | LM size | Quality vs FP32 | Use when |
+|---|---:|---|---|
+| FP32 (upstream) | 640 MB | reference | offline rendering, high-end devices |
+| **FP16 (this repo)** | **336 MB** | **indistinguishable from FP32** | **mobile devices with native fp16** |
+| INT8 ([sister repo](https://huggingface.co/REALBITS/MOSS-TTS-Nano-100M-ONNX-int8)) | 196 MB | ~3 dB envelope shift, slightly more monotone delivery | size-constrained, fastest |
+
+Triple A/B objective analysis (Adam voice, same long demo, seed=42):
+
+| Metric | FP32 | FP16 | INT8 |
+|---|---:|---:|---:|
+| RMS dBFS | -20.79 | -20.49 | -20.18 |
+| F0 mean (Hz) | 110.0 | 110.7 | 114.0 |
+| **F0 std (Hz)** | **32.7** | **35.1** | **21.9** ← more monotone |
+| Voice-band envelope MAE vs FP32 | 0 dB | 9.7 dB ← content-length variance | 13.4 dB |
+| Silent-region noise floor RMS | 0.0014 | 0.0016 | 0.0017 |
+
+## Notes on fp16 numerics
+
+The conversion uses
+`onnxruntime.transformers.float16.convert_float_to_float16(keep_io_types=True)`
+with these ops kept at fp32 (precision-sensitive):
+- `LayerNormalization`
+- `GroupNormalization`
+- `Softmax`
+- `Cast`
+
+This preserves the standard "fp16 forward pass with fp32 norm/softmax" recipe
+that's robust across most transformer architectures. The audio output is
+visually indistinguishable from fp32 in spectral envelope (within the noise
+floor of run-to-run sampling variance — the model has stochastic decoding).
+
+## Speed expectations
+
+| Hardware | FP16 RTF | Notes |
+|---|---:|---|
+| x86 desktop CPU (no AVX-512 BF16) | **~0.75** (slower than fp32 0.38) | ORT-CPU has no fp16 GEMM; expands to fp32 + downconverts |
+| Snapdragon 8 Gen 2/3 ARM cores | TBD | Native fp16 SIMD; expect close to int8 |
+| Apple A14+ | TBD | Native fp16 SIMD |
+| ARM Cortex-X1 / older | varies | Less native support |
+
+If you're targeting x86 CPU, **use the int8 variant instead** — fp16 is a loss there.
+
+## Quick test (Python ORT)
+
+```python
+from onnxruntime import InferenceSession, SessionOptions
+from huggingface_hub import snapshot_download
+
+lm_dir = snapshot_download("REALBITS/MOSS-TTS-Nano-100M-ONNX-fp16")
+codec_dir = snapshot_download(
+    "OpenMOSS-Team/MOSS-Audio-Tokenizer-Nano-ONNX",
+    allow_patterns=["moss_audio_tokenizer_decode_*", "*.json"],
+)
+# Drive prefill → loop {sampler → decode_step} → codec.decode_full
+# (or codec.decode_step for streaming) per upstream onnx_tts_runtime.py.
+```
+
+## On-device example
+
+This repo is a v2 candidate for the
+[Fictures MOSS-TTS-Nano Android prototype](https://github.com/dev-realbits/Fictures/tree/main/mobile/moss-tts-nano-prototype).
+Switch to it via:
+
+```bash
+adb shell setprop debug.mosstts.variant FP16
+```
+
+then restart the app and tap Download.
+
+## How it was quantized
+
+Repro: see
+[`_dryrun/13_quantize_fp16.py`](https://github.com/dev-realbits/Fictures/blob/main/mobile/moss-tts-nano-prototype/_dryrun/13_quantize_fp16.py).
+Three steps:
+1. `onnx.load(load_external_data=True)` for each LM graph
+2. `convert_float_to_float16(keep_io_types=True, op_block_list=[LayerNorm, GroupNorm, Softmax, Cast])`
+3. Save with `save_as_external_data=True`; hash the resulting `.data` files;
+   collapse byte-identical pairs (prefill + decode_step share the underlying
+   transformer weights) into one shared `moss_tts_global_shared_fp16.data`
+   and rewire both graphs' external_data location attributes.
+
+## License + attribution
+
+Apache 2.0, inherited from upstream
+[OpenMOSS/MOSS-TTS-Nano](https://github.com/OpenMOSS/MOSS-TTS-Nano).
+
+## Known limitations
+
+Same as the upstream and int8 variant: English / Chinese / Japanese only;
+no microphone voice cloning out of the box (use the 18 builtin voices in
+`browser_poc_manifest.json`); codec stays fp32.