Загрузка видео...

Не удалось загрузить видео

На главную

ComfyUI-WanVideoWrapper now supports SteadyDancer: like WanAnimate - human image animation framework; produces high-fidelity, coherent motion

wildmindai's profile picture

Wildminder

3,526 subscribers

42,246 просмотров • 8 месяцев назад •via X (Twitter)

ИскусствоОбразованиеНаука и технологии
LIVE

Anya Rossi• Live Now

Private livecam show

Комментарии: 0

Нет доступных комментариев

Здесь появятся комментарии из оригинального поста

Похожие видео

introducing Kling 2.1 this new model produces hyper-realistic motion, supports image inputs, and is super fast! try it now in Krea Video.
0:19

introducing Kling 2.1 this new model produces hyper-realistic motion, supports image inputs, and is super fast! try it now in Krea Video.

KREA AI

72,068 просмотров • 1 год назад

Cinematic Motion Prompts for Fashion The following prompt examples leverage the composition of our starting image to create high-fidelity video motion. Let's go 👇
0:12

Cinematic Motion Prompts for Fashion The following prompt examples leverage the composition of our starting image to create high-fidelity video motion. Let's go 👇

Marco

26,791 просмотров • 7 месяцев назад

LTX-2.3 is now supported in ComfyUI Major quality improvements from Lightricks: - Finer visual details - Better 9:16 portrait videos - Cleaner audio - More stable image-to-video motion - Smarter prompt understanding - Clearer text rendering Examples 🧵👇
0:19

LTX-2.3 is now supported in ComfyUI Major quality improvements from Lightricks: - Finer visual details - Better 9:16 portrait videos - Cleaner audio - More stable image-to-video motion - Smarter prompt understanding - Clearer text rendering Examples 🧵👇

ComfyUI

53,221 просмотров • 5 месяцев назад

Introducing 📦𝗔𝗿𝘁𝗶𝗟𝗮𝘁𝗲𝗻𝘁🔧 (SIGGRAPH Asia 2025) — a high-quality 3D diffusion model that explicitly models object articulation, paving the way for richer, more realistic assets in embodied AI and simulation: – Generates fully articulated 3D objects – Physically plausible joints & motion – High-fidelity 3D Gaussian appearance – Supports generation from a single real image arXiv: Project: Code (coming soon):
0:16

Introducing 📦𝗔𝗿𝘁𝗶𝗟𝗮𝘁𝗲𝗻𝘁🔧 (SIGGRAPH Asia 2025) — a high-quality 3D diffusion model that explicitly models object articulation, paving the way for richer, more realistic assets in embodied AI and simulation: – Generates fully articulated 3D objects – Physically plausible joints & motion – High-fidelity 3D Gaussian appearance – Supports generation from a single real image arXiv: Project: Code (coming soon):

Xingang Pan

11,520 просмотров • 8 месяцев назад

This is complex crowds with Ray3. From packed stadiums to bustling streets, large groups remain coherent. Every subject stays clear and consistent as scenes unfold. High fidelity preserves scale, motion, and detail ensuring realism across even the most complicated scenarios.
0:17

This is complex crowds with Ray3. From packed stadiums to bustling streets, large groups remain coherent. Every subject stays clear and consistent as scenes unfold. High fidelity preserves scale, motion, and detail ensuring realism across even the most complicated scenarios.

Luma AI

407,262 просмотров • 10 месяцев назад

This is preserved anatomy with Ray3. From faces and hands to bodies in motion, every form stays true. Humans and creatures remain clear and coherent in every action. Retain character structure with high fidelity, holding details together even through complex cinematic scenes.
0:17

This is preserved anatomy with Ray3. From faces and hands to bodies in motion, every form stays true. Humans and creatures remain clear and coherent in every action. Retain character structure with high fidelity, holding details together even through complex cinematic scenes.

Luma AI

60,389 просмотров • 10 месяцев назад

🔥 Wan2.2 Animate level up on WaveSpeedAI— now smarter than ever!!!🚀 Get more precise poses, easier controls, and seamless character swaps — just upload one image + one video, and watch your character take the stage! 🎭 Now, it even supports anime-style motion with ultra smooth and expressive flow!!!🤩🤩🤩 🎬Try the upgraded Wan Animate on WaveSpeedAI → #WanAnimate #WaveSpeedAI #AIAnimation #AIvideo #ImageToVideo #AIart #AnimeStyle
0:15

🔥 Wan2.2 Animate level up on WaveSpeedAI— now smarter than ever!!!🚀 Get more precise poses, easier controls, and seamless character swaps — just upload one image + one video, and watch your character take the stage! 🎭 Now, it even supports anime-style motion with ultra smooth and expressive flow!!!🤩🤩🤩 🎬Try the upgraded Wan Animate on WaveSpeedAI → #WanAnimate #WaveSpeedAI #AIAnimation #AIvideo #ImageToVideo #AIart #AnimeStyle

WaveSpeedAI

30,804 просмотров • 9 месяцев назад

🔥 Wan 2.2 Spicy is LIVE! Still using Grok Imagine? This is better. High dynamics. Smooth motion. No limits. The perfect model for generating vivid, unrestricted videos — now with full LoRA support. 🎞️ Try it now: 👉 Image-to-Video: 👉 Image-to-Video + LoRA: #WaveSpeedAI #WAN2_2Spicy #AIVideo #Animation #LoRA
0:19

🔥 Wan 2.2 Spicy is LIVE! Still using Grok Imagine? This is better. High dynamics. Smooth motion. No limits. The perfect model for generating vivid, unrestricted videos — now with full LoRA support. 🎞️ Try it now: 👉 Image-to-Video: 👉 Image-to-Video + LoRA: #WaveSpeedAI #WAN2_2Spicy #AIVideo #Animation #LoRA

WaveSpeedAI

49,419 просмотров • 9 месяцев назад

Stadium-scale, studio-quality motion capture is launching soon… Full 3D control, stable bone lengths and realistic ground-foot contact for high-fidelity animation in Unreal Engine, Blender or any 3D engine. This is not pose estimation—this is dynamic, high-precision, real-world data capturing at AAA quality, providing you with lifelike avatars and seamless animation like never before. 📷⚽️
0:13

Stadium-scale, studio-quality motion capture is launching soon… Full 3D control, stable bone lengths and realistic ground-foot contact for high-fidelity animation in Unreal Engine, Blender or any 3D engine. This is not pose estimation—this is dynamic, high-precision, real-world data capturing at AAA quality, providing you with lifelike avatars and seamless animation like never before. 📷⚽️

Move AI

67,288 просмотров • 1 год назад

One image. One prompt. And this came out. ☕✨ Made this in the Pollo AI app — Image to Video with Seedance 2.0 4K. Smooth motion, nice lighting — looks like a real UGC ad. Honestly, high-end content is way easier now.
0:15

One image. One prompt. And this came out. ☕✨ Made this in the Pollo AI app — Image to Video with Seedance 2.0 4K. Smooth motion, nice lighting — looks like a real UGC ad. Honestly, high-end content is way easier now.

Theo

22,617 просмотров • 23 дней назад

Alisson Pereira's MR2V (Masked Reference-to-Video) -- a new IC LoRA for LTX Video that lets you paint a mask over any part of a clip, feed it a reference image, and inpaint that region with high motion fidelity. Even fast-moving subjects hold up.
0:18

Alisson Pereira's MR2V (Masked Reference-to-Video) -- a new IC LoRA for LTX Video that lets you paint a mask over any part of a clip, feed it a reference image, and inpaint that region with high motion fidelity. Even fast-moving subjects hold up.

Banodoco

22,165 просмотров • 4 месяцев назад

ICYMI: Image generation is now available as a Responses API tool—and includes new capabilities: 🌊 Streaming: see partial images as they load ⤴️ Multi-turn edits: high fidelity edits with cache savings 🌐 Combine with tools like MCP or web search for generations using live data
0:16

ICYMI: Image generation is now available as a Responses API tool—and includes new capabilities: 🌊 Streaming: see partial images as they load ⤴️ Multi-turn edits: high fidelity edits with cache savings 🌐 Combine with tools like MCP or web search for generations using live data

OpenAI Developers

100,922 просмотров • 1 год назад

🇨🇳 Another great Chinese Model, OmniHuman-1.5 from ByteDance Turns 1 image plus a voice track into expressive avatar video by pairing a System 1 and System 2 inspired planner with a Diffusion Transformer, Produces coherent motion for over 1 minute with moving camera and multi character scenes. Most avatar models move to the beat of the audio but miss meaning, so gestures feel generic and emotions feel shallow. The fix here is a Multimodal LLM planner that listens to the speech and drafts a structured plan describing intent, emotions, beats, and high level actions, which gives the motion engine clear semantic targets instead of only rhythm. The motion engine is a Multimodal Diffusion Transformer that fuses the plan with audio, the single reference image, and optional text prompts, then synthesizes continuous body, face, and head motion that matches both words and tone. A key trick is a Pseudo Last Frame, a synthetic target that summarizes the next expected state, which stabilizes fusion across modalities and keeps motion consistent over long spans. From just 1 image and speech, the system outputs speaking avatars with synchronized lips, context aware gestures, and continuous camera movement, and it also supports multi character interactions without manual choreography. Reported results show strong lip sync accuracy, high video quality, natural motion, and close match to text prompts, and the same setup works on nonhuman characters too.
0:18

🇨🇳 Another great Chinese Model, OmniHuman-1.5 from ByteDance Turns 1 image plus a voice track into expressive avatar video by pairing a System 1 and System 2 inspired planner with a Diffusion Transformer, Produces coherent motion for over 1 minute with moving camera and multi character scenes. Most avatar models move to the beat of the audio but miss meaning, so gestures feel generic and emotions feel shallow. The fix here is a Multimodal LLM planner that listens to the speech and drafts a structured plan describing intent, emotions, beats, and high level actions, which gives the motion engine clear semantic targets instead of only rhythm. The motion engine is a Multimodal Diffusion Transformer that fuses the plan with audio, the single reference image, and optional text prompts, then synthesizes continuous body, face, and head motion that matches both words and tone. A key trick is a Pseudo Last Frame, a synthetic target that summarizes the next expected state, which stabilizes fusion across modalities and keeps motion consistent over long spans. From just 1 image and speech, the system outputs speaking avatars with synchronized lips, context aware gestures, and continuous camera movement, and it also supports multi character interactions without manual choreography. Reported results show strong lip sync accuracy, high video quality, natural motion, and close match to text prompts, and the same setup works on nonhuman characters too.

Rohan Paul

63,859 просмотров • 11 месяцев назад

🚨 TRELLIS.2 is now live on fal! 🎯 Image-to-3D model producing up to 1536³ PBR textured assets 🎨 Handles arbitrary topology with rich PBR textures (Base Color, Metallic, Roughness, Alpha) ⚡ 16× spatial compression for efficient, scalable, high-fidelity asset generation
0:11

🚨 TRELLIS.2 is now live on fal! 🎯 Image-to-3D model producing up to 1536³ PBR textured assets 🎨 Handles arbitrary topology with rich PBR textures (Base Color, Metallic, Roughness, Alpha) ⚡ 16× spatial compression for efficient, scalable, high-fidelity asset generation

fal

32,588 просмотров • 7 месяцев назад

This is RICULOUSLY good, TRELLIS 3D Generation model by Microsoft! 🔥 Generate high-quality 3D assets from text or image prompts. Supports various formats like Radiance Fields, 3D Gaussians, and meshes Available for FREE on Hugging Face!
0:14

This is RICULOUSLY good, TRELLIS 3D Generation model by Microsoft! 🔥 Generate high-quality 3D assets from text or image prompts. Supports various formats like Radiance Fields, 3D Gaussians, and meshes Available for FREE on Hugging Face!

Vaibhav (VB) Srivastav

62,455 просмотров • 1 год назад

Hailuo 02 is now on Poe! This new video model stands out for its ability to handle real-world physics and motion where other leading-edge models struggle. It supports text and image prompts, and generates videos that are 6 seconds long. (1/2)
0:17

Hailuo 02 is now on Poe! This new video model stands out for its ability to handle real-world physics and motion where other leading-edge models struggle. It supports text and image prompts, and generates videos that are 6 seconds long. (1/2)

Poe

14,654 просмотров • 1 год назад

Vidu Q3 is now available on Akool Experience next-level AI video creation with Vidu Q3 — a powerful multi-modal model that seamlessly blends high-fidelity visuals with perfectly synchronized audio. Powered by Vidu AI, Q3 shines in: Deep narrative understanding Mastery of complex cinematic language Lifelike, human-feeling motion and emotion If storytelling and cinematic quality matter to you, Vidu Q3 is built for it. Try it now on Akool.
0:11

Vidu Q3 is now available on Akool Experience next-level AI video creation with Vidu Q3 — a powerful multi-modal model that seamlessly blends high-fidelity visuals with perfectly synchronized audio. Powered by Vidu AI, Q3 shines in: Deep narrative understanding Mastery of complex cinematic language Lifelike, human-feeling motion and emotion If storytelling and cinematic quality matter to you, Vidu Q3 is built for it. Try it now on Akool.

Akool Inc

1,528,575 просмотров • 6 месяцев назад

The next frontier in AI video isn't better motion - it's better perspective. What used to require a full production setup, camera rig, location scout, multi-shot choreography - now starts with a single image inside Comfy. Creator seungho__yeo ( IG ) used Hyper POV + Seedance 2 to build a complete cinematic space traversal from one still frame: → Aggressive first-person orbit through the full environment → Fly-through movement that inhabits the space, not just records it → Motion blur and single-take continuity that holds the cinematic feel end to end → No cuts. No rig. One image as the source of truth. This is what perspective design looks like at the new baseline.
0:15

The next frontier in AI video isn't better motion - it's better perspective. What used to require a full production setup, camera rig, location scout, multi-shot choreography - now starts with a single image inside Comfy. Creator seungho__yeo ( IG ) used Hyper POV + Seedance 2 to build a complete cinematic space traversal from one still frame: → Aggressive first-person orbit through the full environment → Fly-through movement that inhabits the space, not just records it → Motion blur and single-take continuity that holds the cinematic feel end to end → No cuts. No rig. One image as the source of truth. This is what perspective design looks like at the new baseline.

ComfyUI

33,964 просмотров • 1 месяц назад

Rerender A Video: Zero-Shot Text-Guided Video-to-Video Translation paper page: Large text-to-image diffusion models have exhibited impressive proficiency in generating high-quality images. However, when applying these models to video domain, ensuring temporal consistency across video frames remains a formidable challenge. This paper proposes a novel zero-shot text-guided video-to-video translation framework to adapt image models to videos. The framework includes two parts: key frame translation and full video translation. The first part uses an adapted diffusion model to generate key frames, with hierarchical cross-frame constraints applied to enforce coherence in shapes, textures and colors. The second part propagates the key frames to other frames with temporal-aware patch matching and frame blending. Our framework achieves global style and local texture temporal consistency at a low cost (without re-training or optimization). The adaptation is compatible with existing image diffusion techniques, allowing our framework to take advantage of them, such as customizing a specific subject with LoRA, and introducing extra spatial guidance with ControlNet. Extensive experimental results demonstrate the effectiveness of our proposed framework over existing methods in rendering high-quality and temporally-coherent videos.
0:17

Rerender A Video: Zero-Shot Text-Guided Video-to-Video Translation paper page: Large text-to-image diffusion models have exhibited impressive proficiency in generating high-quality images. However, when applying these models to video domain, ensuring temporal consistency across video frames remains a formidable challenge. This paper proposes a novel zero-shot text-guided video-to-video translation framework to adapt image models to videos. The framework includes two parts: key frame translation and full video translation. The first part uses an adapted diffusion model to generate key frames, with hierarchical cross-frame constraints applied to enforce coherence in shapes, textures and colors. The second part propagates the key frames to other frames with temporal-aware patch matching and frame blending. Our framework achieves global style and local texture temporal consistency at a low cost (without re-training or optimization). The adaptation is compatible with existing image diffusion techniques, allowing our framework to take advantage of them, such as customizing a specific subject with LoRA, and introducing extra spatial guidance with ControlNet. Extensive experimental results demonstrate the effectiveness of our proposed framework over existing methods in rendering high-quality and temporally-coherent videos.

AK

375,160 просмотров • 3 лет назад