Sources
Loading...
Additional media
Loading...
DeepNewz, mobile.
People-sourced. AI-powered. Unbiased News.
Fun thought exercise for learning more about AI: I've been using AI to help me come up with prompts for AI image generation (on a separate app) I'll explain what I like or don't, and then iterate again. It really helps create renderings that resemble my vision.
Introducing OuteTTS 0.3 1B & 500M 🔥 > Zero shot voice cloning > Multilingual (en, jp, ko, zh, fr, de) > Trained on 20,000 hours of audio > Powered by OLMo-1B & Qwen 2.5 0.5B > Speed & emotion control > Powered by HF grants 🤗 Open science ftw! https://t.co/Tj3LswgQXC
Hailuo MiniMax: Subject Reference I've been experimenting with expressions and emotions using a consistent character and I'm getting impressive results. Prompt below: https://t.co/RH4AiXnUoV
OuteAI has launched OuteTTS 0.3, featuring two models: 1B and 500M. This advanced speech synthesis model supports multilingual capabilities, including English, Japanese, Korean, Chinese, French, and German. It is trained on 20,000 hours of audio and offers zero-shot voice cloning, speed, and emotion control. The model is powered by OLMo-1B and Qwen 2.5 0.5B, and has received funding through Hugging Face grants. Additionally, Hailuo AI has introduced a new text-to-voice feature that also allows for zero-shot voice cloning, with comparisons being made to F5 TTS, a leading open-source solution. Users are experimenting with various AI voice models, including Hailuo AI's T2A-01-HD, which allows for emotional expression in voice cloning.
