Загрузка видео...

Не удалось загрузить видео

Возникла проблема при загрузке этого видео. Это может быть связано с временными проблемами сети или видео может быть недоступно.

На главную

Neuphonic just released NeuDecide, an audio-first tiny decision model at only 43 MB, running on a single CPU thread!! > Speech directly to tool calls (no transcription needed) > Just 46 ms inference on an M3 MacBook > Runs fully offline, even on mobile & Raspberry Pi > 75.5%... show more

steven

2,665 subscribers

18,676 просмотров • 18 дней назад •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

Комментарии: 12

Фото профиля steven

steven18 дней назад

you can try it here: model: repo: blogpost:

Фото профиля avahgaar

avahgaar18 дней назад

>> 75.5% tool accuracy on SLURP >> runs offline on RasPi

Фото профиля ZenGPT

ZenGPT18 дней назад

looks sick bro!

Фото профиля Kumar Deepanshu

Kumar Deepanshu18 дней назад

too good to be true bro.

Фото профиля vel ⋈ 🤌🧦

vel ⋈ 🤌🧦18 дней назад

this is interesting, that low ms and size is really impressive. Your example is directly relevant to me as i am rooting/building my personalized home-ai attached to smart devices. Any key learnings, challenges or observations that stood out to you during build process?

Фото профиля Sulaiman Khan Ghori

Sulaiman Khan Ghori18 дней назад

Sick

Фото профиля Rictus 🍥❁

Rictus 🍥❁18 дней назад

Oh, can I use that on top of Vimium? I had this idea this week.

Фото профиля Kuldeep

Kuldeep18 дней назад

43mb is tiny

Фото профиля selv Marū Atma-Goma

selv Marū Atma-Goma18 дней назад

do you still retain some kind of monitoring?

Фото профиля Dylan D. Mobley

Dylan D. Mobley18 дней назад

Ummm.....yes

Фото профиля Ziggy | Product engineer, web3

Ziggy | Product engineer, web318 дней назад

do you know how it handles the misses at 75.5%, does it ask to confirm or just call the wrong tool?

Фото профиля Berlian Gor

Berlian Gor18 дней назад

How does it compare to Picovoice and Sherpa?

Похожие видео

You don't need a GPU to run text-to-speech anymore. Kyutai released Pocket TTS, a 100 million parameter speech model that runs entirely on CPU. How it performs: > Nearly 6x faster than real time on a MacBook CPU > First audio chunk arrives in about 200 milliseconds > Runs on just 2 CPU cores > Supports voice cloning and 6 languages Setup takes one command: pip install pocket-tts Text-to-speech that used to need a GPU or a paid API now runs locally, on a laptop, in near real time.

You don't need a GPU to run text-to-speech anymore. Kyutai released Pocket TTS, a 100 million parameter speech model that runs entirely on CPU. How it performs: > Nearly 6x faster than real time on a MacBook CPU > First audio chunk arrives in about 200 milliseconds > Runs on just 2 CPU cores > Supports voice cloning and 6 languages Setup takes one command: pip install pocket-tts Text-to-speech that used to need a GPU or a paid API now runs locally, on a laptop, in near real time.

Simplifying AI

14,498 просмотров • 2 месяцев назад

Introducing the next Raspberry Pi. An Ai model that runs on a credit card.

Introducing the next Raspberry Pi. An Ai model that runs on a credit card.

Founders Inc

157,302 просмотров • 1 год назад

DigitalOcean’s model synthesis tool is now available in public preview. 🆕 A new server-side tool: run multiple models on one inference request and get a single synthesized response, no custom orchestration. Just add it to your existing inference request.

DigitalOcean’s model synthesis tool is now available in public preview. 🆕 A new server-side tool: run multiple models on one inference request and get a single synthesized response, no custom orchestration. Just add it to your existing inference request.

DigitalOcean

2,192,331 просмотров • 3 месяцев назад

Managed to get an on-device transcription model running in BlueNote ( Accuracy is hit or miss but it looks pretty cool

Managed to get an on-device transcription model running in BlueNote ( Accuracy is hit or miss but it looks pretty cool

Daniel Kuntz

33,227 просмотров • 2 лет назад

🚨 ElevenLabs just killed every transcription tool on the market. They dropped Scribe v2 Realtime and it's not just another speech-to-text model. This thing processes audio in real-time with zero lag, handles multiple speakers, and actually understands context not just words. Here's how with real examples: Full thread 🧵

🚨 ElevenLabs just killed every transcription tool on the market. They dropped Scribe v2 Realtime and it's not just another speech-to-text model. This thing processes audio in real-time with zero lag, handles multiple speakers, and actually understands context not just words. Here's how with real examples: Full thread 🧵

Hasan Toor

220,789 просмотров • 11 месяцев назад

Working on an accessibility tool for Deaf/HoH VR users named VR-CC to transcribe their audio in a VR overlay! Words are made more transparent when confidence in the transcription is low. Runs locally, no analytics, no subscription service! Even seems to recognize common sounds!

Working on an accessibility tool for Deaf/HoH VR users named VR-CC to transcribe their audio in a VR overlay! Words are made more transparent when confidence in the transcription is low. Runs locally, no analytics, no subscription service! Even seems to recognize common sounds!

Benaclejames

47,272 просмотров • 3 лет назад

Introducing Kitten TTS, a SOTA tiny text-to-speech model - Just 15M parameters - Runs without a GPU - Model size less than 25 MB - Multiple high-quality voices - Ultra-fast - even runs on low-end edge devices Github and HF links below

Introducing Kitten TTS, a SOTA tiny text-to-speech model - Just 15M parameters - Runs without a GPU - Model size less than 25 MB - Multiple high-quality voices - Ultra-fast - even runs on low-end edge devices Github and HF links below

Divam Gupta

349,016 просмотров • 1 год назад

if an agent emits parallel tool calls, pi used to execute them sequentially. funnily enough, only a handful of people complained. welp, just implemented it. caveat: you'll get flicker if you have many parallel tool calls and their outputs are > terminal height. soon to be released.

if an agent emits parallel tool calls, pi used to execute them sequentially. funnily enough, only a handful of people complained. welp, just implemented it. caveat: you'll get flicker if you have many parallel tool calls and their outputs are > terminal height. soon to be released.

Mario Zechner

35,254 просмотров • 7 месяцев назад

GPS-free visual-inertial navigation in real time on Raspberry Pi 5 with a consumer-grade IMU and camera (and this is not even the slowest CPU this runs on!)

GPS-free visual-inertial navigation in real time on Raspberry Pi 5 with a consumer-grade IMU and camera (and this is not even the slowest CPU this runs on!)

Otto Seiskari

264,759 просмотров • 1 год назад

$Someone open-sourced a 66M parameter model that beats ElevenLabs, OpenAI, and Gemini and runs entirely offline on a Raspberry Pi. it's called Supertonic, a text-to-speech model so fast it turns an entire webpage into audio in under 1 second. locally. offline. for free. text-to-speech has lived in the cloud for years. every spoken character was an API call and a fraction of a cent. Supertonic 3 kills that entirely. the model is 99M parameters. ships as a single ONNX file. hits 167× faster than real-time on a laptop CPU. that's ~1,263 characters of speech per second. other open systems sit at 55–287. no cloud. no API. no GPU. → 31 languages, no separate adapters → works inside a browser tab (WebGPU/WASM) → handles phone numbers, currency, dates — no preprocessing → inline tags for laugh and breath → Turns an entire webpage into audio in under 1 second → 44.1kHz studio-grade WAV, no upsampler needed` and on real-world text, "$5.2M", "(212) 555-0142 ext. 402", "30kph”, it's the only one that reads them correctly. ElevenLabs Flash, OpenAI TTS-1, Gemini 2.5 Flash all fail. Supertonic passes. 11.2k stars. 100% Open Source.$

Someone open-sourced a 66M parameter model that beats ElevenLabs, OpenAI, and Gemini and runs entirely offline on a Raspberry Pi. it's called Supertonic, a text-to-speech model so fast it turns an entire webpage into audio in under 1 second. locally. offline. for free. text-to-speech has lived in the cloud for years. every spoken character was an API call and a fraction of a cent. Supertonic 3 kills that entirely. the model is 99M parameters. ships as a single ONNX file. hits 167× faster than real-time on a laptop CPU. that's ~1,263 characters of speech per second. other open systems sit at 55–287. no cloud. no API. no GPU. → 31 languages, no separate adapters → works inside a browser tab (WebGPU/WASM) → handles phone numbers, currency, dates — no preprocessing → inline tags for laugh and breath → Turns an entire webpage into audio in under 1 second → 44.1kHz studio-grade WAV, no upsampler needed` and on real-world text, "$5.2M", "(212) 555-0142 ext. 402", "30kph”, it's the only one that reads them correctly. ElevenLabs Flash, OpenAI TTS-1, Gemini 2.5 Flash all fail. Supertonic passes. 11.2k stars. 100% Open Source.

Superman

30,416 просмотров • 3 месяцев назад

this weekend, i'm finally replacing Claude for Android with a little pi durable thing that runs fully on my phone. building it directly in pi on my phone. and it has subagents :D (not an official earendil product, just a personal project to battle test durable)

this weekend, i'm finally replacing Claude for Android with a little pi durable thing that runs fully on my phone. building it directly in pi on my phone. and it has subagents :D (not an official earendil product, just a personal project to battle test durable)

Mario Zechner

68,066 просмотров • 24 дней назад

Just dropped on HF: Supertonic TTS, a blazing fast speech model. 🤯 > RTF as low as 0.001 on RTX4090 > Runs on-device (no latency, full privacy) > 66M params + 8+ language SDKs > Browser, mobile, edge — it just works. > You can adjust inference steps, batch processing, and other parameters to match your specific needs! > Open-source. Ready to build with!

Just dropped on HF: Supertonic TTS, a blazing fast speech model. 🤯 > RTF as low as 0.001 on RTX4090 > Runs on-device (no latency, full privacy) > 66M params + 8+ language SDKs > Browser, mobile, edge — it just works. > You can adjust inference steps, batch processing, and other parameters to match your specific needs! > Open-source. Ready to build with!

steven

17,784 просмотров • 11 месяцев назад

A perfect coding model for MLX on Apple silicon.. Qwen delivered again. Runs quite fast on an M3 Ultra. Running the 4-bit quantized with mlx-lm:

A perfect coding model for MLX on Apple silicon.. Qwen delivered again. Runs quite fast on an M3 Ultra. Running the 4-bit quantized with mlx-lm:

Awni Hannun

186,831 просмотров • 1 год назад

Can a $175 board power an offline robot voice assistant? 🤖 It’s cool to see Gemma 4 E2B powering the Reachy Mini robot entirely locally on a Raspberry Pi 5. Inference powered by LiteRT, it handles low latency voice, vision (YOLO) and robot movement completely offline. With text generation 2x normal human speech (~300 words per minute) on just 1.5GB of memory, the Pi 5 + Gemma 4 + LiteRT combo is a demonstration of how lean edge AI can be!

Can a $175 board power an offline robot voice assistant? 🤖 It’s cool to see Gemma 4 E2B powering the Reachy Mini robot entirely locally on a Raspberry Pi 5. Inference powered by LiteRT, it handles low latency voice, vision (YOLO) and robot movement completely offline. With text generation 2x normal human speech (~300 words per minute) on just 1.5GB of memory, the Pi 5 + Gemma 4 + LiteRT combo is a demonstration of how lean edge AI can be!

Google Gemma

30,832 просмотров • 2 месяцев назад

Step-3.5-Flash in action on MLX with OpenCode on a single (distributed testing in progress!) M3 Ultra to create a snake game! 🔥 6bit quantization. Perfect tool calling. Fast & powerful coding model! Recommended Inference Settings: Temperature: 1.0 Top-p: 0.95 Top-k: 40 🧵

Step-3.5-Flash in action on MLX with OpenCode on a single (distributed testing in progress!) M3 Ultra to create a snake game! 🔥 6bit quantization. Perfect tool calling. Fast & powerful coding model! Recommended Inference Settings: Temperature: 1.0 Top-p: 0.95 Top-k: 40 🧵

Ivan Fioravanti ᯅ

17,158 просмотров • 8 месяцев назад

the industry hasn't yet internalized how much serverless tool calling will change agent economics watch me run 1000 agents in parallel on a single archil sandbox, with tool calls going to actual Linux microVMs -- all for 30x less than a traditional sandbox it doesn't matter how cheap or expensive the underlying cpus are if you don't waste them while inference is running the cpu capacity crisis will be solved by running agents on right-sized hardware

the industry hasn't yet internalized how much serverless tool calling will change agent economics watch me run 1000 agents in parallel on a single archil sandbox, with tool calls going to actual Linux microVMs -- all for 30x less than a traditional sandbox it doesn't matter how cheap or expensive the underlying cpus are if you don't waste them while inference is running the cpu capacity crisis will be solved by running agents on right-sized hardware

Hunter Leath

13,838 просмотров • 13 дней назад

Yay, llama2.c can now load and inference the Meta released models! :) E.g. here inferencing the smallest 7B model at ~3 tokens/s on 96 OMP threads on a cloud Linux box. Still just CPU, fp32, one single .c file of 500 lines: expecting ~300 tok/s tomorrow :)

Yay, llama2.c can now load and inference the Meta released models! :) E.g. here inferencing the smallest 7B model at ~3 tokens/s on 96 OMP threads on a cloud Linux box. Still just CPU, fp32, one single .c file of 500 lines: expecting ~300 tok/s tomorrow :)

Andrej Karpathy

410,428 просмотров • 3 лет назад

Local transcription running on iPad Pro M1 🔥 Qwen3-ASR-0.6B from Qwen — fully on-device, no cloud, no API calls. Built with our new MLX-Audio-Swift SDK, hitting 25 tok/s at just 1.9GB of RAM. Test audio? Found Marc Lou's old sales call from one of his awesome newsletters sitting on my device. Try it out & drop us a ⭐️

Local transcription running on iPad Pro M1 🔥 Qwen3-ASR-0.6B from Qwen — fully on-device, no cloud, no API calls. Built with our new MLX-Audio-Swift SDK, hitting 25 tok/s at just 1.9GB of RAM. Test audio? Found Marc Lou's old sales call from one of his awesome newsletters sitting on my device. Try it out & drop us a ⭐️

Prince Canuma

81,427 просмотров • 8 месяцев назад