Dataset with phoneme-aligned librilight datasets for voice training at scale.
Dataset has almost 60k hours of english audio in 2TB Tar file, which i have processed with whisper and then aligned with MFA.
I am planning to train a base model for voice processing using this dataset and then finetune it to provide high quality voice, diarization and voice enhancement models. May be you will find it useful too!
https://github.com/ex3ndr/supervoice-dataset/
Dataset has almost 60k hours of english audio in 2TB Tar file, which i have processed with whisper and then aligned with MFA.
I am planning to train a base model for voice processing using this dataset and then finetune it to provide high quality voice, diarization and voice enhancement models. May be you will find it useful too!
https://github.com/ex3ndr/supervoice-dataset/
GitHub
GitHub - ex3ndr/supervoice-librilight-preprocessed: 60k hours of phoneme-aligned audio from audio books
60k hours of phoneme-aligned audio from audio books - ex3ndr/supervoice-librilight-preprocessed
👍4🔥2
One of the biggest news for startups in US: non-compete agreements are no longer valid for anyone in US!
https://x.com/unusual_whales/status/1782853997164130591?s=46&t=A1zE0kfmZyf1QVe55CmQQA
https://x.com/unusual_whales/status/1782853997164130591?s=46&t=A1zE0kfmZyf1QVe55CmQQA
X (formerly Twitter)
unusual_whales (@unusual_whales) on X
BREAKING: The FTC just banned non-compete agreements, per MorePerfectUnion.
The Federal Trade Commission has issued a final rule making it illegal for bosses to make workers sign noncompetes in any scenario, and voiding nearly all existing noncompetes.
The Federal Trade Commission has issued a final rule making it illegal for bosses to make workers sign noncompetes in any scenario, and voiding nearly all existing noncompetes.
😱5👏2👎1🔥1🦄1
The Flutter community on X is super annoying. They smash half-backed solutions and then start to scream how React Native is bad. Last humorous example is the OTA updates, which bypasses appstore review and allows you to update you app instantly. What's the fun part? It will be 100x slower (literaly stated in docs).
And today we got:
https://www.reddit.com/r/FlutterDev/s/l67wYdcGq6
And today we got:
https://www.reddit.com/r/FlutterDev/s/l67wYdcGq6
Reddit
From the FlutterDev community on Reddit: More layoffs for the flutter team 😬
Explore this post and more from the FlutterDev community
😁2👍1🤡1
Looking for product-focused cofounder
I have decided to change my strategy and started to look for a product co-founder, not a business one.
I have figured out that business part can be learned, but product one cannot.
I don’t believe that any venture can figure out what they are actually doing in a first two years, you will be a very lucky to make it sooner. Therefore I will define what I am interested in, not the something too specific.
I think it is possible to launch a new product that focuses only on tech (founders?) form mostly Bay Area. Right now I am working on a mobile app for open source AI wearables. What I have found that it is possible to capture everyday experiences of people and put it to the app. Also right now many people (here) are very open to try new cheap hardware much easier than just an app. I don’t see the need to focus on other markets for next year or two since people a too sceptical right now out of bay’ bubble.
What is everyone in AI in Bay Area are missing is the good product. Customers are ready to pay, a lot of investors are ready to invest. But it seems no one has a taste anymore and all MVPs are in fact too broken, too shallow or does not deliver.
I think community here have a good MVPs, but they all need a strong product.
My app is here: https://getbubble.org
What it is doing is building a feed of your experiences, right now (except basic product things) it is quite expensive to run (~2$ per day per user), but this could be solved.
Unfortunately I can’t focus on both tech and product!
Therefore if someone knows anyone please, share this post with them!
I have decided to change my strategy and started to look for a product co-founder, not a business one.
I have figured out that business part can be learned, but product one cannot.
I don’t believe that any venture can figure out what they are actually doing in a first two years, you will be a very lucky to make it sooner. Therefore I will define what I am interested in, not the something too specific.
I think it is possible to launch a new product that focuses only on tech (founders?) form mostly Bay Area. Right now I am working on a mobile app for open source AI wearables. What I have found that it is possible to capture everyday experiences of people and put it to the app. Also right now many people (here) are very open to try new cheap hardware much easier than just an app. I don’t see the need to focus on other markets for next year or two since people a too sceptical right now out of bay’ bubble.
What is everyone in AI in Bay Area are missing is the good product. Customers are ready to pay, a lot of investors are ready to invest. But it seems no one has a taste anymore and all MVPs are in fact too broken, too shallow or does not deliver.
I think community here have a good MVPs, but they all need a strong product.
My app is here: https://getbubble.org
What it is doing is building a feed of your experiences, right now (except basic product things) it is quite expensive to run (~2$ per day per user), but this could be solved.
Unfortunately I can’t focus on both tech and product!
Therefore if someone knows anyone please, share this post with them!
getbubble.org
Bubble AI
The best mobile app for open Wearable AI
❤8🦄2👍1🥰1
Small sneak peek into what I have built
This is an open source hardware and a bunch of software to run an AI that always on and listens and transcribes everything around you. Then AI agents extract memories and create visualisations.
Right now most of the stuff is running out of my basement simply because otherwise it would cost ~$40 per user.
I am running them of one-two GPUs, one my favourite is RTX 4000 Ada. Which is costly but very low power while providing 20gb of VRam.
Recently it was finally approved by Google and now in both stores I have an app:
https://getbubble.org
This is an open source hardware and a bunch of software to run an AI that always on and listens and transcribes everything around you. Then AI agents extract memories and create visualisations.
Right now most of the stuff is running out of my basement simply because otherwise it would cost ~$40 per user.
I am running them of one-two GPUs, one my favourite is RTX 4000 Ada. Which is costly but very low power while providing 20gb of VRam.
Recently it was finally approved by Google and now in both stores I have an app:
https://getbubble.org
👍8❤3👏2🔥1🥰1💩1
Last weekend we won a big Lllama 3 Hackaton!
We built a smart glasses from scratch using off-the-shelf parts in $20 and it runs entirely on Open Source networks.
This kind of blown up everywhere - now everyone knows what we have done and we have 1500+ of preorders in the first day. Gizmodo took small interview, virtually anyone i am talking two knows about our glasses. I have built firmware and base app (everything up to LLM) and guys benchmarked different llms and picked the best approach.
What we found is that Moondream v2 + Llama 3 is capable to run similar to what OpenAI demonstrated. It can sense what's happening around, answering questions. You can get faster than realtime responses via Groq API for Llama 70B, TTS is also quite fast. STT is also 100x of realtime. Ofc openai is much better at this, but hacked in the weekend solution also works.
My goal getting street creds in bay area is slowly working!
We built a smart glasses from scratch using off-the-shelf parts in $20 and it runs entirely on Open Source networks.
This kind of blown up everywhere - now everyone knows what we have done and we have 1500+ of preorders in the first day. Gizmodo took small interview, virtually anyone i am talking two knows about our glasses. I have built firmware and base app (everything up to LLM) and guys benchmarked different llms and picked the best approach.
What we found is that Moondream v2 + Llama 3 is capable to run similar to what OpenAI demonstrated. It can sense what's happening around, answering questions. You can get faster than realtime responses via Groq API for Llama 70B, TTS is also quite fast. STT is also 100x of realtime. Ofc openai is much better at this, but hacked in the weekend solution also works.
My goal getting street creds in bay area is slowly working!
Gizmodo
Cerebral Valley Hackers Build $20 Open Source Smart Glasses
A five-person team at a San Francisco hackathon created an open source approach to Meta's AI-powered Ray-Bans.
🔥15🦄5❤4
My first audio pre-training finished!
Thankfully my small cluster of 4090 finished pre-training of my audio foundational model! It took about 30 days to finish. This model is akin to original GPT (not chat one) in essence that it is not really that useful without finetuning, but you can finetnue it to do denoising, speech enhancement, speaker separation, text-to-speech tasks. All this tasks are very very important for AI Wearables and every single company i have talked too are struggling with it.
Thankfully my small cluster of 4090 finished pre-training of my audio foundational model! It took about 30 days to finish. This model is akin to original GPT (not chat one) in essence that it is not really that useful without finetuning, but you can finetnue it to do denoising, speech enhancement, speaker separation, text-to-speech tasks. All this tasks are very very important for AI Wearables and every single company i have talked too are struggling with it.
👍6🤔3💅2👨💻1
This media is not supported in your browser
VIEW IN TELEGRAM
SuperVoice Enhance
My new model for speech enhance and denoising. This model processes noisy or damaged audio extracts high quality voice. Should work with most languages out of the box.
Model and sources: https://github.com/ex3ndr/supervoice-enhance
My new model for speech enhance and denoising. This model processes noisy or damaged audio extracts high quality voice. Should work with most languages out of the box.
Model and sources: https://github.com/ex3ndr/supervoice-enhance
🔥14❤2