From f64cd14c367b4fa098c52bd69c5ecb1d791746e9 Mon Sep 17 00:00:00 2001 From: Eason WaveKat Date: Fri, 17 Apr 2026 10:32:59 +1200 Subject: [PATCH 1/3] docs: add model weights section to README Co-Authored-By: Claude Opus 4.6 (1M context) --- README.md | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/README.md b/README.md index 0d6fde8..3d9cffd 100644 --- a/README.md +++ b/README.md @@ -62,6 +62,14 @@ Set `WAVEKAT_MODEL_DIR` to load from a local directory and skip all downloads. All backends produce `AudioFrame<'static>` from [`wavekat-core`](https://github.com/wavekat/wavekat-core) — the same type consumed by `wavekat-vad` and `wavekat-turn`. +## Model weights + +ONNX-converted weights are published under the [`wavekat`](https://huggingface.co/wavekat) organization on Hugging Face. + +| Backend | Repository | Precision | +|---------|------------|-----------| +| Qwen3-TTS | [wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX](https://huggingface.co/wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX) | FP32, INT4 | + ## Architecture ``` From a70535ee3161ab7de5b85ca01a6950d124c56c28 Mon Sep 17 00:00:00 2001 From: Eason WaveKat Date: Fri, 17 Apr 2026 10:36:45 +1200 Subject: [PATCH 2/3] docs: move model weights next to feature flags Co-Authored-By: Claude Opus 4.6 (1M context) --- README.md | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/README.md b/README.md index 3d9cffd..09ac3cd 100644 --- a/README.md +++ b/README.md @@ -62,14 +62,6 @@ Set `WAVEKAT_MODEL_DIR` to load from a local directory and skip all downloads. All backends produce `AudioFrame<'static>` from [`wavekat-core`](https://github.com/wavekat/wavekat-core) — the same type consumed by `wavekat-vad` and `wavekat-turn`. -## Model weights - -ONNX-converted weights are published under the [`wavekat`](https://huggingface.co/wavekat) organization on Hugging Face. - -| Backend | Repository | Precision | -|---------|------------|-----------| -| Qwen3-TTS | [wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX](https://huggingface.co/wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX) | FP32, INT4 | - ## Architecture ``` @@ -110,6 +102,14 @@ _RTF < 1.0 = faster-than-real-time. Lower is better._ _To update: run `make bench-csv-cuda` on target hardware, then commit `bench/results/`._ +## Model weights + +ONNX-converted weights are published under the [`wavekat`](https://huggingface.co/wavekat) organization on Hugging Face. + +| Backend | Repository | Precision | +|---------|------------|-----------| +| Qwen3-TTS | [wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX](https://huggingface.co/wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX) | FP32, INT4 | + ## Feature flags ### Backends From aad42e49775d18a92db5573012473a01808b4776 Mon Sep 17 00:00:00 2001 From: Eason WaveKat Date: Fri, 17 Apr 2026 10:39:44 +1200 Subject: [PATCH 3/3] docs: place model weights right after backends Co-Authored-By: Claude Opus 4.6 (1M context) --- README.md | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) diff --git a/README.md b/README.md index 09ac3cd..94b8bc7 100644 --- a/README.md +++ b/README.md @@ -23,6 +23,14 @@ Same pattern as | [Qwen3-TTS](https://huggingface.co/Qwen/Qwen3-TTS) | `qwen3-tts` | ✅ Available | Apache 2.0 | | [CosyVoice](https://github.com/FunAudioLLM/CosyVoice) | `cosyvoice` | 🚧 Planned | Apache 2.0 | +## Model weights + +ONNX-converted weights are published under the [`wavekat`](https://huggingface.co/wavekat) organization on Hugging Face. + +| Backend | Repository | Precision | +|---------|------------|-----------| +| Qwen3-TTS | [wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX](https://huggingface.co/wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX) | FP32, INT4 | + ## Quick start ```sh @@ -102,14 +110,6 @@ _RTF < 1.0 = faster-than-real-time. Lower is better._ _To update: run `make bench-csv-cuda` on target hardware, then commit `bench/results/`._ -## Model weights - -ONNX-converted weights are published under the [`wavekat`](https://huggingface.co/wavekat) organization on Hugging Face. - -| Backend | Repository | Precision | -|---------|------------|-----------| -| Qwen3-TTS | [wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX](https://huggingface.co/wavekat/Qwen3-TTS-1.7B-VoiceDesign-ONNX) | FP32, INT4 | - ## Feature flags ### Backends