JorgeJGnz

Remember how we used to play as kids, how reality and creativity would mix and we'd create new worlds and stories out of cardboard boxes and kitchen utensils.. how nature around us could turn from a bush in our garden into a dense jungle on an alien planet? As adults, a lot of us forgot that playfulness. But it's still there, waiting to be unleashed :D #art #ai

Watch Starship’s third flight test → spacex.com/launches twitter.com/i/broadcasts/1…

Starship’s hot-stage separation was the first time this technique has been done successfully with a vehicle of this size

Vaaaamoooosss!!!! Un logro impresionante de @PLD_Space y todo su equipo, que consigue un hito histórico! Más que merecido!! Felicidades!! Un honor para mí poder ser parte de esta historia y compartirla! #VamosMiura!

Falta analizar los datos. Pero en cualquier caso todas las personas que trabajan en @PLD_Space pueden estar orgullosísimas: Miura 1 ha volado un perfil de misión completa en su primer lanzamiento. Y eso no pueden decirlo muchas empresas. Han hecho historia. #VamosMIURA!

Just finished my Master's Thesis! 👨‍🎓 Modified a DDPM to support multi-label conditioning with AdaGN for face gen tasks using CelebA-Dialog dataset. Outperformed Stable Diffusion Nano fine-tuned with LoRA as VAE was sub-optimal for 64x64 images #DiffusionModels #StableDiffusion

Can we apply diffusion directly on MLP weights? Yes!!! We show generalization to new 3D shapes and 4D mesh animations: "HyperDiffusion: Generating Implicit Neural Fields with Weight-Space Diffusion" #ICCV2023! Project: ziyaerkoc.com/hyperdiffusion Video: youtu.be/wjFpsKdo-II

LDM3D: Latent Diffusion Model for 3D proposes a Latent Diffusion Model for 3D (LDM3D) that generates both image and depth map data from a given text prompt, allowing users to generate RGBD images from text prompts. The LDM3D model is fine-tuned on a dataset of tuples containing an RGB image, depth map and caption, and validated through extensive experiments. Also develop an application called DepthFusion, which uses the generated RGB images and depth maps to create immersive and interactive 360-degree-view experiences using TouchDesigner paper page: huggingface.co/papers/2305.10…

PER ASPERA AD ASTRA @PLD_Space

Can a diffusion model replicate emergent behaviors when trained for multi-agent simulations? Would it generalize better if trained with a variety of simulations? Btw the dataset is available on HuggingFace bit.ly/40Dwgv4 (4/4)

The big problem here is that doing this requires a specific dataset for each simulation. To avoid doing a complete simulation you have to make hundreds or thousands of complete simulations. Not sure it's really worth it although it seems like an interesting application (3/4)

Diffusion models can act as a low-fidelity short-term simulators. I think this can be useful for "previewing" very expensive simulations: large fluid simulations, complex systems, multi-agents, etc (1/4) #DiffusionModels #PyTorch #GenerativeAI

First fully integrated Starship lifting off for the first time!

🤯 Imagine being back at school and learning about gravity with #QuestPro! 🤯 This scene takes only a few minutes to create: Two drag and drops and just a few clicks on our physics editor. Wishlist today! oculus.com/experiences/qu… #MixedReality #MetaQuest #MetaQuestPro #AR

alright, we're officially living in the future 🤯 this app that lets you relive your memories in VR is just *incredible*

Love the community explorations of ChatGPT, from capabilities (github.com/f/awesome-chat…) to limitations (cookup.ai/chatgpt/where-…). No substitute for the collective power of the internet when it comes to plumbing the uncharted depths of a new deep learning model.

Privacy is hard. link.springer.com/chapter/10.100…

Thought-provocative new paper from @geoffreyhinton: what if we could replace backpropagation with something better?

This paper shows some of our research at Meta Reality Labs in reconstructing a user's pose from only the sensors of the Quest headset using Reinforcement Learning. authors: with Jungdam Won and Yuting Ye paper: arxiv.org/abs/2209.09391 video: youtu.be/CkTHsz6Ldas 1/4👇