Autoencoder Program Deep Learning - Search Videos

AEV2A Sound-Visual Correspondence for the Hand Posture Dataset

AEV2A Sound-Visual Correspondence for the Hand Posture Dataset

757 viewsMar 2, 2019

YouTubeViktor Tóth

Selected as a best paper finalist at #CVPR2026: PixelDiT from NVIDIA ResearchIn most image generation models, a pretrained autoencoder compresses the image before any diffusion happens, causing quality loss that accumulates across the entire pipeline.PixelDiT, or Pixel Diffusion Transformers, removes this step entirely. It's a single-stage model that learns the diffusion process directly in pixel space, end-to-end.

Selected as a best paper finalist at #CVPR2026: PixelDiT from NVIDIA ResearchIn most image generation models, a pretrained autoencoder compresses the image before any diffusion happens, causing quality loss that accumulates across the entire pipeline.PixelDiT, or Pixel Diffusion Transformers, removes this step entirely. It's a single-stage model that learns the diffusion process directly in pixel space, end-to-end.

27.4K views1 month ago

Bad Apple!! but it's learned by a Vector Quantized Variational Autoencoder for reconstruction.

Bad Apple!! but it's learned by a Vector Quantized Variational Autoencoder for reconstruction.

140 views6 months ago

YouTubeComputeIQ

I trained an autoencoder that reconstructs images with zero reconstruction loss. No MSE. No image space supervision. The only signal: "According to you, does your output look like your input through your own eyes?" It works. Blog link, demo and summary 👇

I trained an autoencoder that reconstructs images with zero reconstruction loss. No MSE. No image space supervision. The only signal: "According to you, does your output look like your input through your own eyes?" It works. Blog link, demo and summary 👇

68.5K views1 month ago

Reve 2.0 is here! Ever since we moved away from the latent autoencoder, we are making yet another bet to move away from the pure text-input for image generation!

Reve 2.0 is here! Ever since we moved away from the latent autoencoder, we are making yet another bet to move away from the pure text-input for image generation!

11.9M views1 month ago

x.comSongwei Ge

Neuroscience: In-der-Welt-SeinMost of our intelligence is in the world. The brain learns to capture it on the fly.Lately, almost everyone in the AI community is talking about world models. LeCun is correct that LLMs don't model the world. They only model the statistics of word embeddings in language, i.e., in texts generated by humans.What most AI experts fail to realize about world models is that the human brain does not model the world in the deep learning sense. A deep neural net cannot perce

Neuroscience: In-der-Welt-SeinMost of our intelligence is in the world. The brain learns to capture it on the fly.Lately, almost everyone in the AI community is talking about world models. LeCun is correct that LLMs don't model the world. They only model the statistics of word embeddings in language, i.e., in texts generated by humans.What most AI experts fail to realize about world models is that the human brain does not model the world in the deep learning sense. A deep neural net cannot perce

37.6K views3 weeks ago

How many of the big ideas of the past 15 years of AI are downstream of hardware constraints?The big hardware story over that period is that logic has become way cheaper than data transfer.Stacking huge numbers of matrix multiplies was perfect for this hardware regime, because matrix multiplication is logic-intensive but requires less data transfer. And so we got matmul-heavy deep learning.It's interesting to think about what AI would look like in a world where these costs didn't diverge so much.

How many of the big ideas of the past 15 years of AI are downstream of hardware constraints?The big hardware story over that period is that logic has become way cheaper than data transfer.Stacking huge numbers of matrix multiplies was perfect for this hardware regime, because matrix multiplication is logic-intensive but requires less data transfer. And so we got matmul-heavy deep learning.It's interesting to think about what AI would look like in a world where these costs didn't diverge so much.

10.5K views2 weeks ago

x.comDwarkesh Patel

移动无人车自动路径规划之深度强化学习（Deep Reinforcement Learning）

6K viewsJun 24, 2023

bilibiliMATLAB之智能计算

First Ai tech I’ve seen with real life utility.Tech is gaining traction on GitHub today getting close to 1k stars and multiple ones are getting deployed on pump. I’ll be pushing the first one devved by staccanna team. He is in the community.@imdevPU23 created an AI + IoT plant health platform that combines Realtime environmental sensing from ESP32 (temperature, humidity, light, soil moisture), and Image-based disease detection using a trained deep learning model.The project delivers live dashboa

16.3K views2 weeks ago

Over 500,000 people have used Databricks Free Edition to learn, experiment, and build with data and AI. Today, we're making it a lot more powerful.Free Edition now includes five new products:𝗚𝗲𝗻𝗶𝗲 𝗖𝗼𝗱𝗲 — generates, runs, and iterates on code autonomously on your behalf𝗦𝗲𝗿𝘃𝗲𝗿𝗹𝗲𝘀𝘀 𝗚𝗣𝗨𝘀 — access to GPUs for deep learning, fine-tuning, and inference workloads𝗟𝗮𝗸𝗲𝗯𝗮𝘀𝗲 — a fully managed Postgres-compatible database built for data apps and AI agents𝗔𝗴𝗲𝗻𝘁 𝗕𝗿𝗶𝗰𝗸𝘀

3.5K views3 weeks ago

x.comDatabricks

Deep learning bros and sisters, don't sleep on this.You can cluster millions of documents in embedding space, mass-annotate them, visualize them... basically for free and within seconds.

53.1K views1 month ago

New short course: Fast & Efficient LLM Inference with vLLM, built in partnership with @RedHat and taught by @cedricclyburn.Learn to quantize an open-source LLM, serve it with vLLM, and benchmark your deployment across speed, cost, and accuracy.Free to enroll:

46.1K views1 month ago

x.comDeepLearning.AI

Spent 4 months building the ultimate mosquito killer: an artillery cannon guided by computer vision + deep learning.Trained a custom model to detect and lock onto mosquitoes using a DSLR + zoom lens setup.The dataset collection phase was brutal — the mosquitoes definitely fought back 🦟

69K views1 month ago

x.comSteven Cheng

Yann LeCun ignored every rule in the textbook.And the textbook was wrong.He built neural networks with more parameters than data points.Every statistician alive said that shouldn’t work.LeCun: “Every pre-deep learning textbook told you, you need to have fewer parameters than you have data samples.”He trained them on non-convex objectives with no guarantee of convergence.Every optimization theorist alive said he was out of his mind.LeCun: “All those things you read in textbooks and they tell you

6.5K views1 month ago

Trie Data Structure - GeeksforGeeks

geeksforgeeks.org

In the last 30 days, Microsoft canceled Claude Code for thousands of engineers, Uber burned its entire 2026 AI budget by April, and Nvidia admitted AI costs more than humans.This pattern says something everyone is ignoring.Here is why this matters more than the news cycle is making it sound.For 2 years, every CEO on every earnings call told you the same story.AI will eliminate jobs, cut costs, boost margins. Stock market loved it.Fired workers, stock goes up. Announced AI plans, stock goes up.Th

10.4K views1 month ago

x.comVaibhav Sisinty

🚨BREAKING: THE AI BUBBLE HAS STARTED TO BURST MICROSOFT JUST TOLD 100,000 ENGINEERS TO STOP USING CLAUDE BECAUSE THE BILLS EXPLODED.. UBER BURNED ITS ENTIRE ANNUAL AI BUDGET BY APRIL.. Microsoft invested $5 billion in Anthropic.. gave 100,000 engineers Claude Code access.. encouraged adoption.. watched usage explode.. then the invoices arrived.. and issued an internal order to cancel nearly all Claude Code licenses by end of June and force everyone onto their own cheaper tool..the company that

551.4K views1 month ago

x.comCrypto Jargon

So many people in replies puzzled. "Bro, why not use Process Explorer from sysinternals?" You're missing the point. The time-to-market mentality has clouded your mind.There are people in this world that have passion. Some of them have passion for programming.They love growing their knowledge in how systems work, diving deep into internals, learning how to do things in a software realm so that it's aligned with how hardware works.Some coined the term "mechanical sympathy"People explore their pass

204.2K views1 month ago

x.comŁukasz | Wookash Podcast

See more