Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

Modality Forcing - turns FLUX.2.klein into 3D-aware models. - joint RGB-D, I2D, and D2I synthesis in one model - pixel-space depth tokenization - preserves T2I quality via self-distillation

wildmindai's profile picture

Wildminder

10,373 subscribers

13,049 Aufrufe • vor 17 Tagen •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Kommentare

Keine Kommentare verfügbar

Kommentare vom Original-Post werden hier angezeigt

Ähnliche Videos

Introducing Modality Forcing, a recipe for post-training T2I models for SOTA RGB-Depth generation! Text-to-image (T2I) models learn rich representations of the spatial world. How do we build on this prior for high-quality depth generation? 🧵 [1/6]
0:25

Introducing Modality Forcing, a recipe for post-training T2I models for SOTA RGB-Depth generation! Text-to-image (T2I) models learn rich representations of the spatial world. How do we build on this prior for high-quality depth generation? 🧵 [1/6]

Bardienus Duisterhof

63,883 Aufrufe • vor 20 Tagen

📢 Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation Got only one or a few images and wondering if recovering the 3D environment is a reconstruction or generation problem? Why not do it with a generative reconstruction model! We show that a camera-conditioned video diffusion model can be transformed into a generative reconstruction model that directly outputs a high-quality 3D Gaussian Splatting representation through self-distillation, without requiring real-world training data. Check out our results in the video (wait for dynamic scenes in the second half!) : Project Page: Code and Models: Paper:
0:50

📢 Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation Got only one or a few images and wondering if recovering the 3D environment is a reconstruction or generation problem? Why not do it with a generative reconstruction model! We show that a camera-conditioned video diffusion model can be transformed into a generative reconstruction model that directly outputs a high-quality 3D Gaussian Splatting representation through self-distillation, without requiring real-world training data. Check out our results in the video (wait for dynamic scenes in the second half!) : Project Page: Code and Models: Paper:

Sherwin Bahmani

66,523 Aufrufe • vor 9 Monaten

Generative Novel View Synthesis with 3D-Aware Diffusion Models abs: project page:
0:24

Generative Novel View Synthesis with 3D-Aware Diffusion Models abs: project page:

AK

304,708 Aufrufe • vor 3 Jahren

Most multi-view reconstruction models need full supervision. We show they can self-improve without any ground truth labels. Introducing SelfEvo: Self-Improving 4D Perception via Self-Distillation. Up to +36.5% in video depth, +20.1% in camera estimation, zero annotation.
0:45

Most multi-view reconstruction models need full supervision. We show they can self-improve without any ground truth labels. Introducing SelfEvo: Self-Improving 4D Perception via Self-Distillation. Up to +36.5% in video depth, +20.1% in camera estimation, zero annotation.

Qianqian Wang

24,612 Aufrufe • vor 2 Monaten

FLUX.2 Klein in pure pixel space! No VAE. AsymFlow - hyper-realistic images by working directly in pixel space rather than using compressed latent representations. - sharper textures, superior visual fidelity - 40% faster - low-rank noise parameterization to solve high-dimensional bottlenecks ComfyUI support incoming
0:42

FLUX.2 Klein in pure pixel space! No VAE. AsymFlow - hyper-realistic images by working directly in pixel space rather than using compressed latent representations. - sharper textures, superior visual fidelity - 40% faster - low-rank noise parameterization to solve high-dimensional bottlenecks ComfyUI support incoming

Wildminder

27,210 Aufrufe • vor 1 Monat

FLUX.2 [klein] 4B + 9B are the newest BFL models in the Flux family—combining image generation + image editing in one compact architecture. Built for interactive workflows and quick iteration, with extremely fast inference on distilled variants. Two models. Two modes. ⚡️
0:24

FLUX.2 [klein] 4B + 9B are the newest BFL models in the Flux family—combining image generation + image editing in one compact architecture. Built for interactive workflows and quick iteration, with extremely fast inference on distilled variants. Two models. Two modes. ⚡️

ComfyUI

19,899 Aufrufe • vor 5 Monaten

I went up to Flux 2 Klein 9B param model with 2 steps This model is seriously under rated and will look forward to fine tuning one.
0:35

I went up to Flux 2 Klein 9B param model with 2 steps This model is seriously under rated and will look forward to fine tuning one.

Dev Ed

109,803 Aufrufe • vor 4 Monaten

Diffusion models generate high-quality images but require hundreds of forward passes. MIT CSAIL and Adobe Research introduce Distribution Matching Distillation (DMD), a distillation approach that converts costly multi-step diffusion models into fast one-step generators. A thread 🧵
0:24

Diffusion models generate high-quality images but require hundreds of forward passes. MIT CSAIL and Adobe Research introduce Distribution Matching Distillation (DMD), a distillation approach that converts costly multi-step diffusion models into fast one-step generators. A thread 🧵

MIT CSAIL

34,383 Aufrufe • vor 2 Jahren

Claude distillation has been a big topic this week while I am (coincidentally) writing Chapter 8 on model distillation. In that context, I shared some utilities to generate distillation data from all sorts of open-weight models via OpenRouter and Ollama:
1:40

Claude distillation has been a big topic this week while I am (coincidentally) writing Chapter 8 on model distillation. In that context, I shared some utilities to generate distillation data from all sorts of open-weight models via OpenRouter and Ollama:

Sebastian Raschka

62,458 Aufrufe • vor 4 Monaten

3D-aware Conditional Image Synthesis abs: project page:
0:34

3D-aware Conditional Image Synthesis abs: project page:

AK

301,457 Aufrufe • vor 3 Jahren

Rodin Gen-2 API node, Rodin’s most powerful image-to-3D tool, is now live in ComfyUI! Turn any image into a 3D model with one click. It supports 4× Mesh Quality, T/A Pose, PBR textures, and Multiview to 3D model. 👇 Try it now in ComfyUI!
0:29

Rodin Gen-2 API node, Rodin’s most powerful image-to-3D tool, is now live in ComfyUI! Turn any image into a 3D model with one click. It supports 4× Mesh Quality, T/A Pose, PBR textures, and Multiview to 3D model. 👇 Try it now in ComfyUI!

ComfyUI

19,076 Aufrufe • vor 9 Monaten

Current 3D generative models are slow and low quality. We present GRM, a large-scale model that reconstructs 3D Gaussians in 0.1s and generates high-quality 3D assets from text or single images in a few seconds. Demo: 1/4
0:37

Current 3D generative models are slow and low quality. We present GRM, a large-scale model that reconstructs 3D Gaussians in 0.1s and generates high-quality 3D assets from text or single images in a few seconds. Demo: 1/4

Gordon Wetzstein

19,210 Aufrufe • vor 2 Jahren

Robot Learning needs 4D world models! Robot Learning needs 4D world models! Robot Learning needs 4D world models! We introduce TesserAct, a 4D embodied world model that can simulate how agents interact with the 3D world over time! We achieve this by simply extending a pre-trained 2D video generation model to jointly predict RGB, depth, and surface normals. It enables: 1️⃣ Much better policy learning in the wild 2️⃣ Temporal + spatial coherence in 4D dynamic prediction 3️⃣ Novel view synthesis for embodied scenes Code: Paper Link: Project page:
0:32

Robot Learning needs 4D world models! Robot Learning needs 4D world models! Robot Learning needs 4D world models! We introduce TesserAct, a 4D embodied world model that can simulate how agents interact with the 3D world over time! We achieve this by simply extending a pre-trained 2D video generation model to jointly predict RGB, depth, and surface normals. It enables: 1️⃣ Much better policy learning in the wild 2️⃣ Temporal + spatial coherence in 4D dynamic prediction 3️⃣ Novel view synthesis for embodied scenes Code: Paper Link: Project page:

Chuang Gan

43,265 Aufrufe • vor 1 Jahr

Chinese models ! Llm to 3D ! Connected Deepseek to Flux to Hunyuan 3D ! All inside ComfyUI ! Unlimited of 3D generation model all by itself ! Kartel.ai , ComfyUI
0:20

Chinese models ! Llm to 3D ! Connected Deepseek to Flux to Hunyuan 3D ! All inside ComfyUI ! Unlimited of 3D generation model all by itself ! Kartel.ai , ComfyUI

Lovis Odin

77,779 Aufrufe • vor 1 Jahr

Lucas Roedel shared a tutorial on his free Blender add-on that turns 3D models into 2D pixel art, featuring Pokémon's Tatsugiri as an example. Watch here:
0:41

Lucas Roedel shared a tutorial on his free Blender add-on that turns 3D models into 2D pixel art, featuring Pokémon's Tatsugiri as an example. Watch here:

80 LEVEL

91,323 Aufrufe • vor 9 Monaten

Messing around with the new Flux Controlnet Depth model
0:30

Messing around with the new Flux Controlnet Depth model

Sean

122,365 Aufrufe • vor 1 Jahr

You can now generate, edit, and convert a reference image into 3D inside one tool. Tripo's Image Generation runs Nano Banana and GPT to build the reference, then turns it into a model. Text or photo in, 3D out. One pipeline, zero handoffs.
0:21

You can now generate, edit, and convert a reference image into 3D inside one tool. Tripo's Image Generation runs Nano Banana and GPT to build the reference, then turns it into a model. Text or photo in, 3D out. One pipeline, zero handoffs.

Alvaro Cintas

56,156 Aufrufe • vor 22 Tagen

I had an awesome time interviewing idan shenfeld and Jonas Hübotter from MIT and ETH Zurich about self-distillation. this very promising post-training paradigm where the model acts as its own teacher by conditioning on environment feedback or demonstrations. we cover the SDPO algo for reinforcement learning with rich feedback and SDFT for continual learning without forgetting along with many applications. we dig into how it works, why it's simpler and faster than GRPO, and where this is already showing up in production systems. table of content: 0:00 - what is self distillation 2:50 - idan (MIT) and jonas (ETH Zurich) introduction and motivation 18:40 - different perspective of on-policy self-distillation (presentation) 36:00 - metacognition and specificity in self-distillation 37:24 - very long hard task and self-distillation 42:00 - continual learning with self-distillation (presentation) 1:16:50 - what is next in this research direction? 1:20:00 - is there any experience with subjective feedbacks? 1:22:50 - quality vs number of feedbacks? 1:26:40 - what setting would self-distillation struggle vs GRPO? the slides were super crisp really cool of them to share! enjoy my guys 🌹
1:31:19

I had an awesome time interviewing idan shenfeld and Jonas Hübotter from MIT and ETH Zurich about self-distillation. this very promising post-training paradigm where the model acts as its own teacher by conditioning on environment feedback or demonstrations. we cover the SDPO algo for reinforcement learning with rich feedback and SDFT for continual learning without forgetting along with many applications. we dig into how it works, why it's simpler and faster than GRPO, and where this is already showing up in production systems. table of content: 0:00 - what is self distillation 2:50 - idan (MIT) and jonas (ETH Zurich) introduction and motivation 18:40 - different perspective of on-policy self-distillation (presentation) 36:00 - metacognition and specificity in self-distillation 37:24 - very long hard task and self-distillation 42:00 - continual learning with self-distillation (presentation) 1:16:50 - what is next in this research direction? 1:20:00 - is there any experience with subjective feedbacks? 1:22:50 - quality vs number of feedbacks? 1:26:40 - what setting would self-distillation struggle vs GRPO? the slides were super crisp really cool of them to share! enjoy my guys 🌹

Yacine Mahdid

12,988 Aufrufe • vor 2 Monaten

☕ The new LATTE3D model from #NVIDIAResearch turns text prompts into high-quality 3D representations of objects and animals within a second on a single RTX GPU. #GTC24 👀
0:43

☕ The new LATTE3D model from #NVIDIAResearch turns text prompts into high-quality 3D representations of objects and animals within a second on a single RTX GPU. #GTC24 👀

NVIDIA AI Developer

10,811 Aufrufe • vor 2 Jahren