Orthus.AI’s Post

Did you know that text to speech, or TTS, is not limited to major languages only? Nowadays, there are TTS systems that support many different languages and dialects, making the technology globally useful 🌎️🤩 #TextToSpeech #TTS

I just published "Unbabel’s LLM Tower: How to deploy the model locally". Tower is a multilingual 7B parameter Large Language Model (LLM) optimized for translation-related tasks. It was released by Unbabel in collaboration with the SARDINE Lab at Instituto Superior Técnico, and the MICS lab at CentraleSupélec at the University of Paris-Saclay. This post is a quick guide to deploying Tower locally. #largelanguagemodels #machinetranslation #deployment #llmtower #textgenerationinference

🤔 Machine translation just ask a LLM to do it? While LLMs can translate that doesn't mean they should. What if we could utilize smaller models that were trained to translate between specific languages? What if there was a pipeline that automatically loads models based on the source to target language? 🚀 Enter txtai's translation pipeline! The Translation pipeline automatically detects languages and searches the Hugging Face Hub for the best specialized model to perform the translation. These specialized models are often smaller than LLMs and much faster. Link to code: https://lnkd.in/eefCwRDu

StreamSpeech is an "All in One" seamless model for offline and simultaneous speech recognition, speech translation and speech synthesis. https://lnkd.in/eSYyTTgP #streamspeech #speechrecognition

Large Language Models are changing the game, and with them comes a whole new language. But don't worry, we've got you covered! Think of this blog as your trusty translator, turning those LLM riddles into plain English. Check it out now! #language #LLM #translation #Onix #datametica #google

- LLM routing is a powerful concept that helps manage large language models

𝐏𝐫𝐨𝐣𝐞𝐜𝐭 𝐅𝐨𝐜𝐮𝐬: I'm developing a device that can ease sign language translation without the use of camera-based systems, whose accuracy and feasibility can vary based on a user's access to technology.

An auto-translation feature in DocOrigin allows for the switching of language usage within a single dynamic run. Read the full article: Create Any Document in Any Language ▸ Link in comment #MultiLanguage #CustomerCommunications #Forms

大家好！ We're happy to announce the world's first bilingual Chinese-English embedding model with 8192 token-length. Highlights include: - Bilingual Model: Encodes texts in both English and Chinese, allowing use of either as query or target document - Extended 8K Token-Length: Can process large text passages, exceeding most other open-source models - Compact and Efficient: Just 322MB (161m parameters) and 768 output dimensions; High performance on standard computer hardware w/out GPU, enhancing accessibility https://lnkd.in/eheEMA-k

Confs: IAFPA Annual Conference: We now welcome proposals for presentations in three different categories: oral presentations, poster presentations, and work in progress posters. The conference aims to foster discussions on various speech-related topics, including (but not limited to): Speaker/voice comparison Speech and audio enhancement Transcription Speaker profiling Questioned utterance analysis Authentication of recordings LADO (Language Analysis for the Determination of Origin) Automatic speaker recognition/comparison Vo