Step-Audio-EditX's update is here!
🚀 Key updates：
- Emotion & Style: Huge gains in fidelity → the results speak for themselves.
- More Human: New paralinguistic tags make it feel more alive than ever.
- Better Flow: Super-smooth control over speech rates.
🔥 Open Training: SFT, DPO, and GRPO code is out. Your data, your model!
⚡ Max Efficiency: Powered by vLLM for lightning-fast training and inference.
🤖 How to start :
GitHub: https://t.co/y3nPcjR0N9
Demo page:https://t.co/zbq4bOk2JB
Live Demo:https://t.co/9Jd1tNUZ07
Huggingface: https://t.co/akKNpJq0kl
⭐️Star it, fork it, build with it!
inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
Why it matters
The update ships SFT, DPO and GRPO training code alongside the model, so teams can tune emotion and pacing control on their own speech data instead of accepting the released voice.
article👏🏻Congratulations！Step3-VL-10B was selected for HuggingFace Daily Papers…
postStep-Audio-R1.1 opens weights for a speech model that reasons in real time