正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

I trained a tiny language model to write bad tech jokes and run on a microcontroller 19M parameters 11 MB runs on a conference badge (RP2350), no internet, about 2 tokens a second it learned English from half a billion tokens of tiny kids stories, then I trained it... show more

Sam Meech Ward

4,799 subscribers

16,503 次观看 • 15 天前 •via X (Twitter)

科学技术喜剧

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

I love this opensource project! Someone just put a 28.9M parameter AI model on an ESP32-S3. A chip that costs only a few dollars. No cloud. No API. No internet. It runs locally at ~9.5 tokens/s and can generate stories on a tiny screen. This changes the way we think about AI hardware. The next generation of AI toys, robots, and IoT devices may not need the cloud at all. Small chips are getting smarter. 🤖

I love this opensource project! Someone just put a 28.9M parameter AI model on an ESP32-S3. A chip that costs only a few dollars. No cloud. No API. No internet. It runs locally at ~9.5 tokens/s and can generate stories on a tiny screen. This changes the way we think about AI hardware. The next generation of AI toys, robots, and IoT devices may not need the cloud at all. Small chips are getting smarter. 🤖

Steven Cheng

105,352 次观看 • 3 个月前

Muse gadgets running on a conference badge Alexandr Wang Ported to micropython to run on the RP2350

Muse gadgets running on a conference badge Alexandr Wang Ported to micropython to run on the RP2350

Wes Bos

64,604 次观看 • 1 个月前

I had #ChatGPT write an evocative prompt for #midjourney – then I trained a #stablediffusion model on those images to create this animation. Fractal Diffusion✨

I had #ChatGPT write an evocative prompt for #midjourney – then I trained a #stablediffusion model on those images to create this animation. Fractal Diffusion✨

Gregory Wieber

192,815 次观看 • 3 年前

Fuck yeah! MaskGCT - New open SoTA Text to Speech model! 🔥 > Zero-shot voice cloning > Emotional TTS > Trained on 100K hours of data > Long form synthesis > Variable speed synthesis > Bilingual - Chinese & English > Available on Hugging Face Fully non-autoregressive architecture: > Stage 1: Predicts semantic tokens from text, using tokens extracted from a speech self-supervised learning (SSL) model > Stage 2: Predicts acoustic tokens conditioned on the semantic tokens. Synthesised: "Would you guys personally like to have a fake fireplace, an electric one, in your house? Or would you rather have a real fireplace? Let me know down below. Okay everybody, that's all for today's video and I hope you guys learned a bunch of furniture vocabulary!" TTS scene keeps getting lit! 🐐

Fuck yeah! MaskGCT - New open SoTA Text to Speech model! 🔥 > Zero-shot voice cloning > Emotional TTS > Trained on 100K hours of data > Long form synthesis > Variable speed synthesis > Bilingual - Chinese & English > Available on Hugging Face Fully non-autoregressive architecture: > Stage 1: Predicts semantic tokens from text, using tokens extracted from a speech self-supervised learning (SSL) model > Stage 2: Predicts acoustic tokens conditioned on the semantic tokens. Synthesised: "Would you guys personally like to have a fake fireplace, an electric one, in your house? Or would you rather have a real fireplace? Let me know down below. Okay everybody, that's all for today's video and I hope you guys learned a bunch of furniture vocabulary!" TTS scene keeps getting lit! 🐐

Vaibhav (VB) Srivastav

139,160 次观看 • 2 年前

Nobody asked about it a Tiny Desk Fany on a green screen Take my tiny beautiful biscuit and put her on your desk

Nobody asked about it a Tiny Desk Fany on a green screen Take my tiny beautiful biscuit and put her on your desk

Foxdee

137,137 次观看 • 1 年前

We took a 30B model and split it in two to write tokens in parallel instead of one at a time. Introducing Nemotron-Labs-TwoTower: a diffusion language model from NVIDIA Research adapted from Nemotron-3-Nano-30B-A3B. Here’s how it works: one half holds the context, the other writes the tokens, with both reusing the pretrained model instead of training a new one from scratch. We found it kept 98.7% of the original model’s quality at 2.42× faster generation.

We took a 30B model and split it in two to write tokens in parallel instead of one at a time. Introducing Nemotron-Labs-TwoTower: a diffusion language model from NVIDIA Research adapted from Nemotron-3-Nano-30B-A3B. Here’s how it works: one half holds the context, the other writes the tokens, with both reusing the pretrained model instead of training a new one from scratch. We found it kept 98.7% of the original model’s quality at 2.42× faster generation.

NVIDIA AI

763,589 次观看 • 4 个月前

You can now vibe code a language model. From a single prompt, GPT‑5.6 built the entire training pipeline and trained a model from scratch on my iMessage history. Locally on my Mac. It now generates replies in my writing style.

You can now vibe code a language model. From a single prompt, GPT‑5.6 built the entire training pipeline and trained a model from scratch on my iMessage history. Locally on my Mac. It now generates replies in my writing style.

Pietro Schirano

722,551 次观看 • 3 个月前

A small insect bit me and left a tiny mark on my hand. After a while, it started looking like this. What should I do about it? 🥲

A small insect bit me and left a tiny mark on my hand. After a while, it started looking like this. What should I do about it? 🥲

B✞g Dav✞es 🗽

4,291,470 次观看 • 2 个月前

we trained a fruit fly to understand emotion We created a neural network from the male drosophila connectome, then trained it on human emotions. It was shocking to see how good it was at identification (better than some humans I know)

we trained a fruit fly to understand emotion We created a neural network from the male drosophila connectome, then trained it on human emotions. It was shocking to see how good it was at identification (better than some humans I know)

Nathan Roll

613,864 次观看 • 1 个月前

The first natively trained 1-bit model: BitNet 2B. Trained on 4 trillion tokens. that can run on CPUs like Apple M2 Native 1.58-bit weights and 8bit activations W158A8 Outperforms LLaMA &close to Qwen 2.5 1.5B in while using only 0.4GB memory versus 2GB and processes tokens 40%

The first natively trained 1-bit model: BitNet 2B. Trained on 4 trillion tokens. that can run on CPUs like Apple M2 Native 1.58-bit weights and 8bit activations W158A8 Outperforms LLaMA &close to Qwen 2.5 1.5B in while using only 0.4GB memory versus 2GB and processes tokens 40%

Md Ismail Šojal 🕷️

43,807 次观看 • 8 个月前

STARBOY perceives its environment through a camera, microphone, temperature sensor, and accelerometer. We trained multiple tiny image models that run locally on the device, letting it recognize human faces and hand gestures.

STARBOY perceives its environment through a camera, microphone, temperature sensor, and accelerometer. We trained multiple tiny image models that run locally on the device, letting it recognize human faces and hand gestures.

Daniel Kuntz

116,785 次观看 • 7 个月前

I built gpu-pii, inspired by Shu's gpu-lexer. A small model, trained from scratch, runs on WebGPU in your browser to detect personal information in any text

I built gpu-pii, inspired by Shu's gpu-lexer. A small model, trained from scratch, runs on WebGPU in your browser to detect personal information in any text

Duncan

10,488 次观看 • 1 个月前

I made my indie game fit on a floppy and run on MS-DOS! (It also runs on Windows 11 and PS5 and everything in between)

I made my indie game fit on a floppy and run on MS-DOS! (It also runs on Windows 11 and PS5 and everything in between)

Sos "Wishlist Lazy Kickers" Sosowski

23,107 次观看 • 7 个月前

VERCEL + CLOUDFLARE + CLAUDE CODE all picked up Jev in 6 days... an AI that can't write a single word became the fastest adopted model on Vercel's AI Gateway Jev decides → big model thinks → code authorizes TypeSafe dropped it on September 15, founder Diogo Almeida worked on ChatGPT research at OpenAI it answers three ways: pick one option, rate it on a scale, or yes/no with a percentage "which agent next?" comes back as Researcher 71% / Coder 18% / Reviewer 11% Vercel says almost 13% of its paying teams tried it within the first 24 hours on Vercel's leaderboard it pulled about 26% of all requests but only 1.8% of tokens TypeSafe's own tests: up to 194x faster and 445x cheaper than regular models TypeSafe calls it a System One Model, named after Kahneman's Thinking, Fast and Slow your expensive model stops burning tokens on yes or no questions and only gets the hard ones 6 days, 110 community use cases, zero words written ↓

VERCEL + CLOUDFLARE + CLAUDE CODE all picked up Jev in 6 days... an AI that can't write a single word became the fastest adopted model on Vercel's AI Gateway Jev decides → big model thinks → code authorizes TypeSafe dropped it on September 15, founder Diogo Almeida worked on ChatGPT research at OpenAI it answers three ways: pick one option, rate it on a scale, or yes/no with a percentage "which agent next?" comes back as Researcher 71% / Coder 18% / Reviewer 11% Vercel says almost 13% of its paying teams tried it within the first 24 hours on Vercel's leaderboard it pulled about 26% of all requests but only 1.8% of tokens TypeSafe's own tests: up to 194x faster and 445x cheaper than regular models TypeSafe calls it a System One Model, named after Kahneman's Thinking, Fast and Slow your expensive model stops burning tokens on yes or no questions and only gets the hard ones 6 days, 110 community use cases, zero words written ↓

leopardracer

16,184 次观看 • 1 个月前

Just dropped on HF: kani-tts-370m A lightweight open-source text-to-speech model that sounds great and runs fast! > 370M parameters — efficient and deployable on consumer GPUs > NanoCodec + LFM2-350M > Natural & expressive voice trained with modern neural TTS techniques > Fast inference: real-time on a single RTX 3060

Just dropped on HF: kani-tts-370m A lightweight open-source text-to-speech model that sounds great and runs fast! > 370M parameters — efficient and deployable on consumer GPUs > NanoCodec + LFM2-350M > Natural & expressive voice trained with modern neural TTS techniques > Fast inference: real-time on a single RTX 3060

steven

34,558 次观看 • 1 年前

On this day, I eat 6 eggs, 2 steaks and a banana. I also trained chest

On this day, I eat 6 eggs, 2 steaks and a banana. I also trained chest

Arran Aro

22,964 次观看 • 1 年前

Bello! Introducing Minionese LM, a language model trained from scratch on my Mac mini by 5.6 Sol Based on Minionese, the language spoken exclusively by Minions (until now)

Bello! Introducing Minionese LM, a language model trained from scratch on my Mac mini by 5.6 Sol Based on Minionese, the language spoken exclusively by Minions (until now)

Maddie D. Reese

95,661 次观看 • 3 个月前

I posted about this on YT a while back, but my main project atm is a dumb little cartoon thing loosely inspired by Classic Internet Culture™ Here's a tiny clip that makes complete sense both with and without context, just to prove it exists.

I posted about this on YT a while back, but my main project atm is a dumb little cartoon thing loosely inspired by Classic Internet Culture™ Here's a tiny clip that makes complete sense both with and without context, just to prove it exists.

🌝NDMytp🌝

14,623 次观看 • 8 个月前

Part 2 of my Hindi film lessons - Chupke Chupke this weekend. Jokes are hard in a new language. I’m watching out for Amitabh Bachchan’s lessons on botany and Dharmendra’s take on English. What else?

Part 2 of my Hindi film lessons - Chupke Chupke this weekend. Jokes are hard in a new language. I’m watching out for Amitabh Bachchan’s lessons on botany and Dharmendra’s take on English. What else?

Alex Ellis

150,671 次观看 • 3 年前