diff --git a/public/authors/julian-m-kleber.jpg b/public/authors/julian-m-kleber.jpg index 12c2272..4934679 100644 Binary files a/public/authors/julian-m-kleber.jpg and b/public/authors/julian-m-kleber.jpg differ diff --git a/public/posts/13-02-24-semester-end/banner.jpg b/public/posts/13-02-24-semester-end/banner.jpg new file mode 100644 index 0000000..527a6cc Binary files /dev/null and b/public/posts/13-02-24-semester-end/banner.jpg differ diff --git a/public/posts/13-02-24-voicebox/banner.png b/public/posts/13-02-24-voicebox/banner.png new file mode 100644 index 0000000..88f7649 Binary files /dev/null and b/public/posts/13-02-24-voicebox/banner.png differ diff --git a/public/posts/13-02-24-voicebox/ogImage.png b/public/posts/13-02-24-voicebox/ogImage.png new file mode 100644 index 0000000..19cfad3 Binary files /dev/null and b/public/posts/13-02-24-voicebox/ogImage.png differ diff --git a/src/pages/blog/13-02-2024-Voicebox.md b/src/pages/blog/13-02-2024-Voicebox.md new file mode 100644 index 0000000..44dcf05 --- /dev/null +++ b/src/pages/blog/13-02-2024-Voicebox.md @@ -0,0 +1,30 @@ +--- +layout: ../../layouts/BlogLayout.astro +slug: 13-02-24-voicebox +title: Voicebox - Text-Guided Multilingual Universal Speech Generation at Scale +description: + Voicebox - Text-Guided Multilingual Universal Speech Generation at Scale +authors: + - Julian M Kleber +format: paper +tags: + - transcription + - audio + - reading group + - music + - sheet music +ogImage: + src: /posts/13-02-24-voicebox/ogImage.png + alt: Paper Image +banner: + src: /posts/13-02-24-voicebox/banner.png + alt: Paper Title +published: 02/13/2024 +--- + +Large-scale generative models such as GPT and DALL-E have revolutionized the research community. These models not only generate high fidelity outputs, but are also generalists which can solve tasks not explicitly taught. In contrast, speech generative models are still primitive in terms of scale and task generalization. In this paper, we present Voicebox, the most versatile text-guided generative model for speech at scale. Voicebox is a non-autoregressive flow-matching model trained to infill speech, given audio context and text, trained on over 50K hours of speech that are not filtered or enhanced. Similar to GPT, Voicebox can perform many different tasks through in-context learning, but is more flexible as it can also condition on future context. Voicebox can be used for mono or cross-lingual zero-shot text-to-speech synthesis, noise removal, content editing, style conversion, and diverse sample generation. In particular, Voicebox outperforms the state-of-the-art zero-shot TTS model VALL-E on both intelligibility (5.9% vs 1.9% word error rates) and audio similarity (0.580 vs 0.681) while being up to 20 times faster. + +Link to paper: [https://arxiv.org/abs/2306.15687](https://arxiv.org/abs/2306.15687 + +[Banner](https://arxiv.org/abs/2111.03017) +[Picture](https://vectorportal.com/de/vector/jukebox-machine.ai/15161) diff --git a/src/pages/blog/13-02-2024-semester-end.md b/src/pages/blog/13-02-2024-semester-end.md new file mode 100644 index 0000000..4fe528a --- /dev/null +++ b/src/pages/blog/13-02-2024-semester-end.md @@ -0,0 +1,28 @@ +--- +layout: ../../layouts/BlogLayout.astro +slug: 13-02-24-semester-end +title: Community Event - BLISS in SS24 - feedback phase and plans for the next semester +description: + Community Event - BLISS in SS24 - feedback phase and plans for the next semester +authors: + - Julian M Kleber +format: event +tags: + - BLISS + - organization + - Community Event +ogImage: + src: /posts/13-02-24-semester-end/banner.jpg + alt: Paper Image +banner: + src: /posts/13-02-24-semester-end/banner.jpg + alt: Paper Title +published: 02/13/2024 +--- + +During the event we will do a brainstorming session about possible future directions of bliss, and planning of the lecture free period. + +Link to paper: [https://arxiv.org/abs/2306.15687](https://arxiv.org/abs/2306.15687 + +[Banner](https://commons.wikimedia.org/wiki/File:Das_letzte_Abendmahl,_mit_Liniendivergenz_der_Perspektive_und_Versatz_in_der_R%C3%BCckwand.jpg) +[Picture](https://commons.wikimedia.org/wiki/File:Das_letzte_Abendmahl,_mit_Liniendivergenz_der_Perspektive_und_Versatz_in_der_R%C3%BCckwand.jpg)