Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

H3 generates 10s in under 3 minutes 🤯 After adding LoRA, the generation time dropped from 14 minutes to just under 3 minutes. God damn! Stack: - H100, 80GB - ComfyUI latest - minimax_h3_fl2va_pruned_bf16 (40GB) - Kijai/MiniMax-H3_comfy on Hugging Face - PyTorch 2.13 + CUDA 13.0 The video quality... show more

DΞV

4,603 subscribers

45,226 görüntüleme • 14 gün önce •via X (Twitter)

Bilim & Teknoloji

Anya Rossi• Live Now

Private livecam show

0 Yorum

Yorum bulunmuyor

Orijinal gönderinin yorumları burada görünecek

Benzer Videolar

This video is just under 3 minutes but well worth a watch. There are some uncomfortable truths spoken, but truths all the same. 📽️Niceboysinthecountry(ig)

This video is just under 3 minutes but well worth a watch. There are some uncomfortable truths spoken, but truths all the same. 📽️Niceboysinthecountry(ig)

🎩Laird of the Manor🎩

13,116 görüntüleme • 9 ay önce

Google just dropped Veo 3 Fast, and it's 5x cheaper than the Quality version! So I made this AI video in 20 minutes and for under $5. Prompts + how-to, here 🧵👇

Google just dropped Veo 3 Fast, and it's 5x cheaper than the Quality version! So I made this AI video in 20 minutes and for under $5. Prompts + how-to, here 🧵👇

TechHalla

250,188 görüntüleme • 1 yıl önce

MiniMax H3 is the SOTA open weight model... but the large model comes with slow gens (depending on your hardware). You can be waiting for minutes just for the generation to miss the mark Here's a ComfyUI workflow to get the most out of H3 without the wait. → H3 Turbo LoRA (4–8 steps) → Model Preview Override + Tiny AutoEncoder H3 for a fast composition preview → Bad preview? stop it, tweak the prompt or seed → Save your precious gpu time Workflow and model links below 👇

MiniMax H3 is the SOTA open weight model... but the large model comes with slow gens (depending on your hardware). You can be waiting for minutes just for the generation to miss the mark Here's a ComfyUI workflow to get the most out of H3 without the wait. → H3 Turbo LoRA (4–8 steps) → Model Preview Override + Tiny AutoEncoder H3 for a fast composition preview → Bad preview? stop it, tweak the prompt or seed → Save your precious gpu time Workflow and model links below 👇

rob - comfyui

26,990 görüntüleme • 11 gün önce

This speech by Josh Frydenberg passionately nails everything the Jewish community is feeling, not just in Australia but around the world. Country after country has failed the Jewish people and chose to gaslight them instead of addressing the obvious root of the problem. I edited silences from his speech to fit under 3 minutes for social media, but everyone REALLY needs to watch the full thing. (first reply in thread)

This speech by Josh Frydenberg passionately nails everything the Jewish community is feeling, not just in Australia but around the world. Country after country has failed the Jewish people and chose to gaslight them instead of addressing the obvious root of the problem. I edited silences from his speech to fit under 3 minutes for social media, but everyone REALLY needs to watch the full thing. (first reply in thread)

The Mossad: Satirical and Awesome

14,060 görüntüleme • 8 ay önce

She makes these recipes seem so easy. I'm making this. Plus, she is really funny. Annette Freckles Easy Mexican Rice Instructions: Cook on high for 10 minutes with 3 cups of water. Turn off the heat but leave the pan on the burner for 5 minutes. Remove the pan from the heat for 10 minutes. DO NOT EVER TAKE OFF THE LID. See recipe in video.

She makes these recipes seem so easy. I'm making this. Plus, she is really funny. Annette Freckles Easy Mexican Rice Instructions: Cook on high for 10 minutes with 3 cups of water. Turn off the heat but leave the pan on the burner for 5 minutes. Remove the pan from the heat for 10 minutes. DO NOT EVER TAKE OFF THE LID. See recipe in video.

DS | Wild Bloom 🌿

78,113 görüntüleme • 5 ay önce

You bought the Stocks on the Ghana Stock Exchange market But is your name actually on them? 👀. Most investors don’t know how their investments are really held. Do you? The CSD explained in under 3 minutes. This is the one video every investor needs to watch. Pooled (Omnibus) or direct model; two ways to hold your investments. Which one are you on? 🤔 #GSE #Stocks #Shares #GhanaStocks

You bought the Stocks on the Ghana Stock Exchange market But is your name actually on them? 👀. Most investors don’t know how their investments are really held. Do you? The CSD explained in under 3 minutes. This is the one video every investor needs to watch. Pooled (Omnibus) or direct model; two ways to hold your investments. Which one are you on? 🤔 #GSE #Stocks #Shares #GhanaStocks

Sage Kenny 🐺🦍 | Digital Marketer

69,316 görüntüleme • 5 ay önce

Nick Wright COOKED James Harden for 3 STRAIGHT MINUTES 😭🤣 “I would not trade Darius Garland for James Harden under ANY circumstances… All this guy does is quit on teams and FOLD in the Playoffs… He’s an all time great player, an all time great quitter, and an all time great choker” 💀 (via First Things First)

Nick Wright COOKED James Harden for 3 STRAIGHT MINUTES 😭🤣 “I would not trade Darius Garland for James Harden under ANY circumstances… All this guy does is quit on teams and FOLD in the Playoffs… He’s an all time great player, an all time great quitter, and an all time great choker” 💀 (via First Things First)

Hater Report

909,122 görüntüleme • 6 ay önce

here's how Strava can come onchain within minutes using Starkzap explained under 5minutes. the video covers everything from > the current problem on strava. > steps taken by the move-to-earn industry > introducing Stravazap essentially Strava with Starkzap > features of Stravazap and how it works > benefits for both the users and for Strava as a company all that an app needs now to come on chain in minutes is just simply using Starkzap.

here's how Strava can come onchain within minutes using Starkzap explained under 5minutes. the video covers everything from > the current problem on strava. > steps taken by the move-to-earn industry > introducing Stravazap essentially Strava with Starkzap > features of Stravazap and how it works > benefits for both the users and for Strava as a company all that an app needs now to come on chain in minutes is just simply using Starkzap.

Brother Akash 🥷

15,654 görüntüleme • 5 ay önce

Christopher Nolan reveals filming with IMAX originally meant waiting 25 minutes after every 3 minutes of footage "These cameras only can shoot just under 3 minutes of film" "The first time I shot IMAX on The Dark Knight, the IMAX techs told us it took 25 minutes to reload the camera and we're like, 'Well, that's not going to work.'" "We now have it to about a minute and a half" "Hoyte is a very big fan of Formula 1, so I think maybe he had his crew watching Formula 1 races because they were on it"

Christopher Nolan reveals filming with IMAX originally meant waiting 25 minutes after every 3 minutes of footage "These cameras only can shoot just under 3 minutes of film" "The first time I shot IMAX on The Dark Knight, the IMAX techs told us it took 25 minutes to reload the camera and we're like, 'Well, that's not going to work.'" "We now have it to about a minute and a half" "Hoyte is a very big fan of Formula 1, so I think maybe he had his crew watching Formula 1 races because they were on it"

JayJay⛓️

599,470 görüntüleme • 27 gün önce

SITUATION EXPLAINED: MiniMax released H3 with open weights, and it's already #1 in video editing. • Priced at $7.80 per minute of 2K video, against Seedance 2.0 at $22.45 and Kling 3.0 at $20.16 • It ranks #1 on Artificial Analysis for video editing, #2 for text-to-video behind Gemini Omni Flash, and #3 for image-to-video • H3 takes text, images, video, and audio in a single context and generates 5-15 second clips at native 2K, 24fps, with native stereo audio • MiniMax is releasing the weights under its Community License within days, which would make it the strongest open-weights video model by a wide margin • It launched the same week ByteDance rolled out Seedance 2.5, which now generates 30-second 4K video from a single prompt • Seedance 2.5's face consistency across a full 30 seconds is the visible jump, earlier models drifted into a different face partway through • Voices are still the obvious tell, the video reads as real long before the audio does • Smaller labs building purpose-built video models are now beating frontier labs at this specific task Theo Jaffee: "This is gonna be one of the biggest examples of things that people are gonna do with these models, historical simulations, basically. Nostalgia videos for zoomers."

SITUATION EXPLAINED: MiniMax released H3 with open weights, and it's already #1 in video editing. • Priced at $7.80 per minute of 2K video, against Seedance 2.0 at $22.45 and Kling 3.0 at $20.16 • It ranks #1 on Artificial Analysis for video editing, #2 for text-to-video behind Gemini Omni Flash, and #3 for image-to-video • H3 takes text, images, video, and audio in a single context and generates 5-15 second clips at native 2K, 24fps, with native stereo audio • MiniMax is releasing the weights under its Community License within days, which would make it the strongest open-weights video model by a wide margin • It launched the same week ByteDance rolled out Seedance 2.5, which now generates 30-second 4K video from a single prompt • Seedance 2.5's face consistency across a full 30 seconds is the visible jump, earlier models drifted into a different face partway through • Voices are still the obvious tell, the video reads as real long before the audio does • Smaller labs building purpose-built video models are now beating frontier labs at this specific task Theo Jaffee: "This is gonna be one of the biggest examples of things that people are gonna do with these models, historical simulations, basically. Nostalgia videos for zoomers."

MTS

55,187 görüntüleme • 22 gün önce

VIDEO | Malviya Nagar fire: AAP leader Somnath Bharti says, "Yesterday, it was reported that the fire brigade took about an hour to an hour and a half to arrive. The fire station that was built during my tenure, next to the Malviya Nagar Metro Station, is quite close to this location. I wanted to measure how much time it should actually take to travel from there to here. So, I just measured it myself using a stopwatch. It took exactly two and a half minutes. That means during the daytime, when there is traffic, it took only two and a half minutes. At 8 am, when traffic is lighter, it would take barely two minutes. So my question is, if the distance can be covered in two minutes, how did it take an hour or an hour and a half for the fire brigade to arrive? This is a question that the BJP's four-engine government cannot avoid answering. The fire brigade is under your responsibility. The MCD is under your responsibility. The Delhi government is under your responsibility. At every level, it is your administration." (Full video available on PTI Videos -

VIDEO | Malviya Nagar fire: AAP leader Somnath Bharti says, "Yesterday, it was reported that the fire brigade took about an hour to an hour and a half to arrive. The fire station that was built during my tenure, next to the Malviya Nagar Metro Station, is quite close to this location. I wanted to measure how much time it should actually take to travel from there to here. So, I just measured it myself using a stopwatch. It took exactly two and a half minutes. That means during the daytime, when there is traffic, it took only two and a half minutes. At 8 am, when traffic is lighter, it would take barely two minutes. So my question is, if the distance can be covered in two minutes, how did it take an hour or an hour and a half for the fire brigade to arrive? This is a question that the BJP's four-engine government cannot avoid answering. The fire brigade is under your responsibility. The MCD is under your responsibility. The Delhi government is under your responsibility. At every level, it is your administration." (Full video available on PTI Videos -

Press Trust of India

36,461 görüntüleme • 2 ay önce

here's how to make ultra realistic AI podcasts: - generate first frames with nano banana pro (prompt is key here) - upscale images - animate with VEO 3 that's it. it literally takes just a few minutes to go from ‘slop’ to ultra-realistic ai videos you can fix voices with minimax or in post production and it'll look indistinguishable from a real video the only difference is the prompt and the workflow

here's how to make ultra realistic AI podcasts: - generate first frames with nano banana pro (prompt is key here) - upscale images - animate with VEO 3 that's it. it literally takes just a few minutes to go from ‘slop’ to ultra-realistic ai videos you can fix voices with minimax or in post production and it'll look indistinguishable from a real video the only difference is the prompt and the workflow

MAX

16,530 görüntüleme • 6 ay önce

‼️EXPOSED: US Marshals Were On Scene At UVU Directing People Out Of The Amphitheatre Within ~5 MINUTES Of Charlie Being Hit😱 This video shows U.S. Marshals arriving and immediately taking control of the area — clearing people out of the amphitheatre just minutes after the hit. That timeline is impossible under normal circumstances. The closest U.S. Marshals office is at the United States Courthouse in Salt Lake City — roughly 45 miles away. Under normal driving conditions that trip takes 45–55 minutes. There is no realistic way deputies from that office arrive in approximately 5 minutes unless they were already staged in the immediate area before the event started. When the response time doesn’t match the distance, it stops looking like a standard reaction and starts looking like prior positioning. Just like the FBI’s impossibly fast arrival, this raises the same unanswered question: why were federal personnel close enough to be on scene that quickly? Tag Candace Owens & RT this. Shoutout to KK for providing me with this never before seen video. FOLLOW Her. She is doing great research! P.S. What else do you notice in this video? i have also added a slowed down version. Drop your thoughts below.

‼️EXPOSED: US Marshals Were On Scene At UVU Directing People Out Of The Amphitheatre Within ~5 MINUTES Of Charlie Being Hit😱 This video shows U.S. Marshals arriving and immediately taking control of the area — clearing people out of the amphitheatre just minutes after the hit. That timeline is impossible under normal circumstances. The closest U.S. Marshals office is at the United States Courthouse in Salt Lake City — roughly 45 miles away. Under normal driving conditions that trip takes 45–55 minutes. There is no realistic way deputies from that office arrive in approximately 5 minutes unless they were already staged in the immediate area before the event started. When the response time doesn’t match the distance, it stops looking like a standard reaction and starts looking like prior positioning. Just like the FBI’s impossibly fast arrival, this raises the same unanswered question: why were federal personnel close enough to be on scene that quickly? Tag Candace Owens & RT this. Shoutout to KK for providing me with this never before seen video. FOLLOW Her. She is doing great research! P.S. What else do you notice in this video? i have also added a slowed down version. Drop your thoughts below.

Project Constitution

102,793 görüntüleme • 20 gün önce

Everybody speak about consistent character for video generation as Minimax, Kling etc... But nothing looking super stable. And so I trained for the first time a LoRA model on Hunyuan video model (open source) on a real person. It take approximatly 15h of training on a A100 GPU with 50 videos. the result is pretty good !! With more training it will allow to have even better result. The result was generated inside ComfyUI. All this video have been generated in AI. Check at , we are doing cool stuff :)

Everybody speak about consistent character for video generation as Minimax, Kling etc... But nothing looking super stable. And so I trained for the first time a LoRA model on Hunyuan video model (open source) on a real person. It take approximatly 15h of training on a A100 GPU with 50 videos. the result is pretty good !! With more training it will allow to have even better result. The result was generated inside ComfyUI. All this video have been generated in AI. Check at , we are doing cool stuff :)

Lovis Odin

93,560 görüntüleme • 1 yıl önce

Black Forest Labs just announced FLUX 3: a unified multimodal model for image, video, audio and action prediction. Founded in Freiburg, Germany, one of the few globally significant European companies in the AI sector. The current release is gated early access, not open source. But if FLUX 3 Dev actually ships with usable open weights, this could become one of the most important open-model releases in generative video yet. Especially after the Chinese Minimax H3 release. BFL claims, based on preliminary internal comparisons, that FLUX 3 outperformed models such as Runway Gen-4.5, Luma Ray 3.2, Kling 3 Pro, and Seedance 2.0. However, these figures stem from the early-access phase and do not constitute independent validation. Be that as it may, it's great to see the quality that today's video models can produce at affordable prices. Something that would have been unthinkable a year ago. Really cool release!

Black Forest Labs just announced FLUX 3: a unified multimodal model for image, video, audio and action prediction. Founded in Freiburg, Germany, one of the few globally significant European companies in the AI sector. The current release is gated early access, not open source. But if FLUX 3 Dev actually ships with usable open weights, this could become one of the most important open-model releases in generative video yet. Especially after the Chinese Minimax H3 release. BFL claims, based on preliminary internal comparisons, that FLUX 3 outperformed models such as Runway Gen-4.5, Luma Ray 3.2, Kling 3 Pro, and Seedance 2.0. However, these figures stem from the early-access phase and do not constitute independent validation. Be that as it may, it's great to see the quality that today's video models can produce at affordable prices. Something that would have been unthinkable a year ago. Really cool release!

Chubby♨️

39,869 görüntüleme • 18 gün önce

My son is interested in Stop Motion video, but all the apps on the store are garbage, too finicky for a 5 year old. I always wanted to just make a decent one, but never had the time. I did it with meʟ in 20 minutes and now he's making Sonic the hedgehog clips nonstop.

My son is interested in Stop Motion video, but all the apps on the store are garbage, too finicky for a 5 year old. I always wanted to just make a decent one, but never had the time. I did it with meʟ in 20 minutes and now he's making Sonic the hedgehog clips nonstop.

Michael Feldstein

182,004 görüntüleme • 1 yıl önce

the video is live. full app built in under 3 hours. raw. uncut. bugs and all. the stack: → Idea Browser — idea to full PRD in minutes → Mobbin — production grade UI patterns from pre-existing apps → — built full stack with plain english (i specialize in frontend but you can also use superbase) what you'll see: HOUR 1: PRD generation (user flows, features, edge cases) HOUR 2: reference hunting (real apps > dribbble fantasy) HOUR 3: full build (frontend and you can use superbase for backend) this isn't a polished tutorial. it's the actual workflow. live. you'll see where it breaks. you'll see how I fix it. the gap between "idea" and "shipped product" is now 3 hours. watch it. then go build something. like + RT if this changes how your new workflow looks like..

the video is live. full app built in under 3 hours. raw. uncut. bugs and all. the stack: → Idea Browser — idea to full PRD in minutes → Mobbin — production grade UI patterns from pre-existing apps → — built full stack with plain english (i specialize in frontend but you can also use superbase) what you'll see: HOUR 1: PRD generation (user flows, features, edge cases) HOUR 2: reference hunting (real apps > dribbble fantasy) HOUR 3: full build (frontend and you can use superbase for backend) this isn't a polished tutorial. it's the actual workflow. live. you'll see where it breaks. you'll see how I fix it. the gap between "idea" and "shipped product" is now 3 hours. watch it. then go build something. like + RT if this changes how your new workflow looks like..

MotionViz

21,716 görüntüleme • 7 ay önce