Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
A Flow Matching-based Text-to-Speech Model with Emoji-driven Style Control
| Date | Stars |
|---|---|
| 2026-07-24 | 1043 |
| 2026-07-25 | 1045 |
| 2026-07-28 | 1060 |
| 2026-07-30 | 1064 |
| 2026-08-06 | 1064 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Irodori-TTS [](https://huggingface.co/Aratako/Irodori-TTS-500M-v3) [](https://huggingface.co/Aratako/Irodori-TTS-600M-v3-VoiceDesign) [](https://huggingface.co/spaces/Aratako/Irodori-TTS-500M-v3-Demo) [](https://huggingface.co/spaces/Aratako/Irodori-TTS-600M-v3-VoiceDesign-Demo) [](LICENSE) Training and inference code for **Irodori-TTS**, a Flow Matching-based Text-to-Speech model. The architecture and training design largely follow [Echo-TTS](https://jordandarefsky.com/blog/2025/echo/), using [DACVAE](https://github.com/facebookresearch/dacvae) continuous latents as the generation target. For an OpenAI-compatible inference API server, see [Irodori-TTS-Server](https://github.com/Aratako/Irodori-TTS-Server). > [!IMPORTANT] > `main` tracks the **v3** codebase and is intended for use with the **Irodori-TTS-500M-v3** base model release. > It also supports the **Irodori-TTS-600M-v3-VoiceDesign** 3-branch VoiceDesign release. > The current code remains backward-compatible with **Irodori-TTS-500M-v2** checkpoints, including **Irodori-TTS-500M-v2-VoiceDesign**. > If you need the previous v2 codebase state, use the `v2` tag. If you need the previous v1 code, use the `v1` tag. > v1 checkpoints / preprocessing are not compatible with v2/v3. > The previous public v1 model is available at [Aratako/Irodori-TTS-500M](https://huggingface.co/Aratako/Irodori-TTS-500M). For model weights and audio samples, please refer to the [base model card](https://huggingface.co/Aratako/Irodori-TTS-500M-v3) and the [VoiceDesign model card](https://huggingface.co/Aratako/Irodori-TTS-600M-v3-VoiceDesign). ## Features - **Flow Matching TTS**: Rectified Flow Diffusion Transformer (RF-DiT) over continuous DACVAE latents - **Voice Cloning**: Zero-shot voice cloning from reference audio - **Multi-modal Voice Design**: v3 VoiceDesign can combine text, reference speech, and caption text for voice identity plus style/emotion control - **Emoji-based Style Control**: Emoji annotations in input text can influence delivery and non-verbal vocal expressions in supported checkpoints - **Automatic Duration Prediction**: v3 base and v3 VoiceDesign checkpoints estimate output length without manual `--seconds` - **Automatic Watermarking**: Generated audio is watermarked with [SilentCipher](https://github.com/sony/silentcipher) when available - **Multi-GPU Training**: Distributed training via `uv run --no-sync torchrun` with gradient accumulation, mixed precision (bf16), and W&B logging - **PEFT LoRA Fine-Tuning**: Parameter-efficient adaptation with PEFT/LoRA for released checkpoints - **Speaker Inversion**: Learn reusable speaker embedding tokens for a target voice while freezing the base model - **Flexible Inference**: CLI, Gradio Web UI, and HuggingFace Hub checkpoint support ## Architecture The current codebase supports two closely related checkpoint families: 1. **Base model (`Aratako/Irodori-TTS-500M-v3`)**: Text encoder + reference latent encoder + diffusion transformer + duration predictor. The reference latent encoder consumes patched DACVAE latents from reference audio for speaker/style conditioning. v2 base checkpoints remain supported for inference. 2. **VoiceDesign model (`Aratako/Irodori-TTS-600M-v3-VoiceDesign`)**: Text encoder + reference latent encoder + caption encoder + diffusion transformer + duration predictor. The v3 VoiceDesign path supports 3-branch conditioning from text, reference speech, and caption text. v2 VoiceDesign remains supported as a backward-compatible caption-only checkpoint family. Shared building blocks: 1. **Text Encoder**: Token embeddings initialized fr
Excerpt of 24,999 characters
Read on GitHub
Japan
27
2
2
1
1
1
Would you bet a product on this? Bounded 0–100 and slow moving.
matched fp:9bd46ba941aca1dd, topic:text-to-speech, topic:tts, topic:voice-cloning