We are thrilled to announce "automatic gradient descent"---a neural network optimiser without hyperparameters. AGD trains out-of-the-box and at ImageNet scale. paper: arxiv.org/abs/2304.05187 PyTorch: github.com/jxbz/agd 1/5

We're live at #AWE2023 with #Monado, your #OpenSource, cross-platfrom XR runtime! Come see it in action at booth 930, and see how our team can help you leverage #OpenXR! #AR #VR #XR #AugmentedReality #VirtualReality #MixedReality AWE Monado OpenXR

Pushing the frontiers of identifying and filtering labeling errors in machine learning datasets. Check out the blog post for more details👇

Machine learning is constantly advancing, it can be hard to keep track. I’ve created a handy daily newsletter in audio format - perfect for your commute mix. Tune in at podcast.mlcult.org

Machine Learning: Translating languages, transcribing voices, and summarizing content in the blink of an eye! The future of seamless communication and information consumption is here. col.la/mlscript #machinelearning #transcription #translation #opensource marcus

Unlock the Power of Words with col.la/mlscript Ever wished for a tool that turns hours of audio into concise summaries? At Collabora, we worked on a completely open source transcription, summarization, and translation pipeline. #machinelearning #transcription Vineet

Yesterday I had the pleasure to talk with Christoph Schuhmann from LAION about my github.com/collabora/Whis… model and the future of Open Source emotive text to speech. Big thanks to Collabora for making this work possible. youtube.com/watch?v=1OBvf3…

Democratizing AI just dropped a new podcast episode: "Open Source Text-To-Speech Projects: WhisperSpeech - In Depth Discussion". Tune in to unpack the evolution of #WhisperSpeech in the #OpenSource world! ➡️col.la/mltext #machinelearning Jakub Piotr Cłapa

📢 New blog post! Meet the latest player in the text-to-speech field - WhisperSpeech, a proudly #opensource model: col.la/wstext #MachineLearning Jakub Piotr Cłapa

Coming up next week: the inaugural edition of #AIdotDev! Join us as Jakub Piotr Cłapa dives into findings from WhisperSpeech, a new #OpenSource text-to-speech model based entirely on properly licensed speech datasets & unrestricted code col.la/aidotdev #AI #ML Jakub Piotr Cłapa

#AI news: By creating a real-time #AIChatbot communication system using fully #OpenSource tools, Collabora's engineers have addressed the unnatural delay in current bot interactions for seamless conversation! Meet #WhisperFusion: openfir.st/whisperfusion #ML #TextToSpeech

A company working on the AI assistant latency problem. Users will hate talking to AIs if they take 20 seconds to answer.

Collabora releases WhisperSpeech demo: huggingface.co/spaces/collabo… Open Source text-to-speech system built by Collabora and LAION by inverting Whisper.

I am on holiday while Collabora’s Open Source WhisperSpeech TTS model landed on the top spot of Spaces this week on Hugging Face ! Thanks a lot AK for helping to spread the word 🚀 Meanwhile, v2 already started brewing… :)

While many companies have built their #machinelearning analysis framework around #GStreamer, no one had made the effort to contribute upstream, until now. Introducing the @GStreamer Analytics Metadata Framework: col.la/gstanalytics #OpenSource #DataAnalytics #ComputerVision

BUD-E on Germany's biggest IT-Blog :) heise.de/news/Bud-E-Lai…

🔥 WhisperFusion: low latency conversations with an AI chatbot incorporating multiple optimization strategies such as torch.compile, TensorRT, Batching, Quantization, KV caching, Attention, Decoding, etc: 🐙 github.com/collabora/Whis… 📹 youtube.com/watch?v=_PnaP0… Collabora 🙏👇

Open Source honored at Embedded World: Collabora's #WhisperFusion has been nominated for Embedded Award 2024! col.la/ea24 #OpenSource #AI #ML #TextToSpeech #AIChatbot

This #EarthDay we're taking a moment to highlight the incredible work that our environmental partners do daily. As 1% for the Planet members, we support them to secure a future for people and our planet.🌎➡️col.la/ed24 #onepercentfortheplanet