i need a way to segment portions of video to alter some and keep the rest as the original..so far i only see comfyui can do this for video well..any idea of another platfrom ..preferably webui to use sam2 with inpainting/controlnet?..my laptop cant run comfy..any feedback woudl be amazing!..thank you for your vid btw
@TechCindy24 күн бұрын
Stable Diffusion webui plugin for controlnet segmentation, etc. (but seems that only for images): github.com/continue-revolution/sd-webui-segment-anything Projects that segment videos, if you have your own GPUs: github.com/achalddave/segment-any-moving github.com/gaomingqi/track-anything I don't have experience segmenting videos myself so just listing some resources here. Hope this helps a bit & best luck!
@srinu57120 күн бұрын
@seanknowles7987 , I am also working for a same use case, I need to delete the some frames from video and keep the rest as the original, I need to remove frames based on the user prompt, it is almost similar to your use case, did you find any models that can fulfill this requirement. ?
@StupitVoltMainАй бұрын
I mean this is actually cool idea, but however there still should be hard-coded scripts. AI and scripts of the talented writers will come hand in hand together
@AlexindiegamedevАй бұрын
I don't necessarily hate AI, I hate when AI takes the jobs from real artists just so corporations can buy their expensive cars
@TechCindyАй бұрын
^^ I'm concerned if AI widens the gap between the rich and the poor.
@AlexindiegamedevАй бұрын
As both a indie game dev working on my own project & someone who plays games, I don't want AI in my games.
@kennethlingadАй бұрын
i observed and figured this out recently, i thought it was just me lol
@TechCindyАй бұрын
Haha nice! We're somewhat connected now.
@ethanhe42Ай бұрын
Where to try this llama3.1?
@TechCindyАй бұрын
There's this platform called Meta AI and you can run llama 3.1 online :)
@JFrogyАй бұрын
I notice your voice seems a bit robotic here. Was it compressed or tuned in any way? Its hard to listen to for a long duration
@TechCindyАй бұрын
I used CapCut to speed it up a little bit. Yes, I feel the sound quality dropped too :( Any suggestions?
@JFrogyАй бұрын
@@TechCindy I think in Audacity you can change the tempo of a audio track and it will speed it up without changing the pitch, but thats just what I saw online. thanks for the vids! :)
@TechCindyАй бұрын
@@JFrogy Wow thank you so much! I'm gonna try it.
@chewmynoodlesАй бұрын
@@TechCindy Please do not speed up your voice. It gets really distracting from an otherwise good video. Just play it regularly and people can speed up the video if they want through youtube playback controllers. Thank you for the video :)
@TechCindyАй бұрын
@@chewmynoodles Thank you for your previous advice! <3
@JFrogyАй бұрын
I could totally see a open world game utilize AI language model for the main characters or companions in a game
@TechCindyАй бұрын
Awesome! Thanks for sharing.
@Thebuzzer1Ай бұрын
Wow amazing video. This is my area of interest. AI and gaming has me hooked.
@TechCindyАй бұрын
Thx my friend! I'm into AI and gaming too, and I'll make more videos.
@huccicoderАй бұрын
Very good video. I like the part about "NPC dont necessarily need advanced AI"!
@TechCindyАй бұрын
Yes yes! Thank you.
@TechCindy2 ай бұрын
Text version available here: tiancaixinxin.medium.com/is-ai-the-future-of-npcs-an-overview-of-npc-intelligence-in-games-71328f12f83c
@DerWisser2 ай бұрын
idk, how I got to your Video but you did a great Job! The quality and the info are on point, keep it up!
@TechCindy2 ай бұрын
Thank you my new friend, I will <3
@extrachannelyt65282 ай бұрын
Does "Enhance prompt" make the loading longer?
@TechCindy2 ай бұрын
I didn't observe a big difference - each video takes a minute or so to generate regardless. The option likely expands your prompt to a slightly longer one with more details.
@extrachannelyt65282 ай бұрын
@@TechCindy i discovered for myself, that uploading a photo with no description and enhance prompt disabled just cause for photo to be zoomed in slowly with no difference
@TechCindy2 ай бұрын
@@extrachannelyt6528 Cool, you made it! I think it's a bug and I'd expect to get moving frames even without a prompt description.
@michaelmita25882 ай бұрын
Hey I tried the demo, but it doesn't seem to work ? It doesnt seem to load anything just says "Generation Error"
@ThugSheep2 ай бұрын
I meam since isabelle is a dog , the second one seems more close fhan the first 😂
@TechCindy2 ай бұрын
haha Isabelle is a human character in Animal Crossing, not a dog 😂
@melodieangelique3385Ай бұрын
@@TechCindy She...is a dog...just a humanoid dog...
@voodoochild420ai3 ай бұрын
pretty cool, it still looks very blurry and not great but im sure they will improve in the future and be very useful
@TechCindy3 ай бұрын
Yeah, hopefully :)
@voodoochild420ai3 ай бұрын
interesting video, cool tip to try and turn off that feature could have better resutls
@3dcool9633 ай бұрын
Hello, how do I import the exported BVH into the BIP skeleton of 3DMAX? I tried to import it, but the result was that there were unrecognizable tracks! Thank you
@TechCindy3 ай бұрын
Follow Dr. Jim Fan on Twitter: x.com/DrJimFan Check out GenAI Summit SF: x.com/genaisummitsf MineDojo: developer.nvidia.com/blog/building-generally-capable-ai-agents-with-minedojo Voyager: github.com/MineDojo/Voyager MetaMorph: metamorph-iclr.github.io/site Eureka: blogs.nvidia.com/blog/eureka-robotics-research Dr. Eureka: eureka-research.github.io/dr-eureka
@dogamer2383 ай бұрын
how do i import the bvh file in unreal engine ?
@TechCindy3 ай бұрын
Hi! People like to use Blender to convert the bvh to fbx and import the fbx to Unreal Engine.
@dogamer2383 ай бұрын
@@TechCindy thanks
@ethanhe423 ай бұрын
amzing video
@ItsCoderDan4 ай бұрын
we are cooked
@BlueprintBro3 ай бұрын
Deep fried
@ericdrob4 ай бұрын
nice. what's the name of the voice actor used at 1:25
@TechCindy4 ай бұрын
I used CapCut's text-to-voice AI.
@prathameshdinkar29664 ай бұрын
Nicely explained! Keep the good work going! Though you have less subscribers, this type of content will soon attract many 😃
@TechCindy4 ай бұрын
Thank you so much! I've just started running a channel and have a lot to learn <3 What content would you like, how to use the AI tools or how to train AI models? Suggestions are welcomed.
@prathameshdinkar29664 ай бұрын
@@TechCindy I think you con consider what you like and good at, though I like how to use AI tools!
@TechCindy4 ай бұрын
Aw you're right. I'll try to make more useful content :)
@TechCindy4 ай бұрын
To try out AnimationGPT- Demo: www.animationgpt.net/enUS/demo GitHub: github.com/fyyakaxyy/animationGPT Video transcript: tiancaixinxin.medium.com/animationgpt-an-ai-to-generate-actions-for-game-characters-6d10d7e7387d
@rtluo15464 ай бұрын
So why do you use the genmo the most, given that genmo is losing under all experiments.
@TechCindy4 ай бұрын
Good question. For a realistic style, Runway > Genmo > Pika. Given that Runway has a limited free trial, I use Genmo the most considering benefit vs. cost :)
@Rolandfart4 ай бұрын
I think that openAI could easily create a model specifically for text to 3D by applying what they know about Sora. Sora clearly is able to create 3D scenes, all that needs to change is the output method.
@TechCindy4 ай бұрын
Yup, we'll see :)
@cosmicfissure9244 ай бұрын
Excelente video Maestra, buen día.
@TechCindy4 ай бұрын
¡Gracias!
@parallaxoldchannel4 ай бұрын
"Can OpenAI Sora solve the problem of text to 3D?" It's not a problem, it's an industry with REAL artists that create 3D Models. We don't need anything "solving" our jobs. The only moment in which it becomes a "problem" that needs "solving" is when people who don't want to work want to produce. So, they don't want to invest time on learning how to 3D Model or developing 3D Models or art for that matter. They see it as something in the way of fulfilling their instant gratifications and thus think of it as a "problem" that needs "solving".
@TechCindy4 ай бұрын
My other video talks about why I think AI can't replace 3D artists. However I think my word choice here is overall appropriate. Is 1+1 a math problem? Yes. Do we have problem with it? No.
@parallaxoldchannel4 ай бұрын
@@TechCindy 1+1 is an equation, a formula with means to an end. If we are looking at the void of a viewport in any 3d software package as a "problem" then the solution is the 3D Modeler aka 3D Artist. The work involved is not a problem that needs solving. It's an equation with means to an end. What you are theorizing is an alternative workflow.
@xitcix83604 ай бұрын
Think of where it will be in the future. AI will be better than humans at making videos. We will have simulations of the world that we can interact with. Imagine how incredible music and movies will be when they're 2X better than humans. Plus, movies and shows are controlled by greedy shareholders, this would give everyone the ability to make anything without restrictions from corporate greed
@parallaxoldchannel4 ай бұрын
@@xitcix8360 Having ambition, drive and the will to learn and create exists without "corporate greed".
@parallaxoldchannel4 ай бұрын
@@xitcix8360 By your logic, your life up until the implementation of current AI...has been 2x less enjoyable.
@ScarletEternity-m2r4 ай бұрын
This is crazy we could have games that build themselves in the future Like imagin playing gta and driving to another state and it just generates the world on the fly
@TechCindy4 ай бұрын
Love that you mentioned gta! I was thinking the other day, if there will be more games in a unlimited space in a realistic art style because of AI.
@TechCindy4 ай бұрын
Text version available here: tiancaixinxin.medium.com/can-openai-sora-solve-3d-modeling-i-tried-this-47c1576a4e8d Mentioned works: [NeRF] kzbin.info/www/bejne/gKaraGx7bdeXgcU [COLMAP] github.com/colmap/colmap [SceneDreamer] scene-dreamer.github.io
@MokuroYamsei5 ай бұрын
Do you know about AGI? What it is?
@_ben_sisko5 ай бұрын
Absolutely nothing that man said makes any sense. AI is going to make the poor even more poor, because most of the jobs they can get will be done by AI, and as AI and robotics get more sophisticated, human labor will become irrelevant. The cost of things going down or even becoming free has nothing to do with how poor people are.
@TechCindy5 ай бұрын
Although many people will lose jobs, I think AI can't replace human creativity. It's so unique and valuable.
@_ben_sisko5 ай бұрын
@@TechCindy What the fuck does that have to do with if people are poor or not?
@Kim-Yo-jong5 ай бұрын
Siqi is a delusional neoliberal.
@TechCindy5 ай бұрын
Thanks for sharing your thoughts.
@XavierNV5 ай бұрын
You lost me with the Tesla claims
@TechCindy5 ай бұрын
I mainly mentioned Tesla's two AI products: self driving & humanoid robots. Lmk how I can explain things further & thanks for the feedback :)
@XavierNV5 ай бұрын
I meant that you lost credibility with the Tesla AI claims, as the demonstrations performed by the humanoid robots have largely been discredited, and the AI systems are not novel
@TechCindy5 ай бұрын
I respect your view. People may have divisive thoughts about an experimental product. Yann LeCun dislikes ChatGPT and I guess it's still fair to say ChatGPT can do a lot of things. Optimus can, for example, pick up eggs, which involves many vision and mechanical problems to solve.
@TechCindy5 ай бұрын
Text version available here: tiancaixinxin.medium.com/big-techs-ai-strategies-each-in-a-phrase-1b9d5ad43ebe
@rafaelrcamargo5 ай бұрын
great video!
@TechCindy5 ай бұрын
Thanks!
@OMGanger5 ай бұрын
You forgot about Mbodi ahem definitely not a shameless plug. You say big tech has every advantage but anyone can download from hf and innovate. There is much more to the game than scale.
@TechCindy5 ай бұрын
My heart always with the individual builders who are brave to innovate. In the past year, I have seen many sad stories, which is why I wanna point out some difficulties and wrong paths to avoid 😢
@tqboy5 ай бұрын
cooooool🤣
@TechCindy5 ай бұрын
Hope you enjoyed it :D
@guozhang87305 ай бұрын
so cool
@qubit9665 ай бұрын
English subtitles would be very helpful.
@TechCindy5 ай бұрын
All my past videos should have English subtitles. Have you found the subtitle?
@qubit9665 ай бұрын
@@TechCindy ah ok I see. Thank you. My mistake. Great content, by the way. Keep up the good work!!!
@TechCindy5 ай бұрын
@@qubit966 Awesome, thanks!!
@TechCindy5 ай бұрын
Text version here: tiancaixinxin.medium.com/2024-ai-quick-learning-guide-43312dac1ebb
@Telencephelon5 ай бұрын
Cool channel. I subscribed
@TechCindy5 ай бұрын
Thank you <3
@Telencephelon5 ай бұрын
Awesome content. Keep it coming. Adjust the volume on your voice or equalize the video audio.
@TechCindy5 ай бұрын
Thanks a lot for the suggestion! Yes, I realized the problem and am doing it for future videos.
@ethanhe425 ай бұрын
Thanks for sharing!
@TechCindy5 ай бұрын
Thanks for the great talk!
@Abhay-pundora5 ай бұрын
Thanks for this video ..
@TechCindy5 ай бұрын
Hope it helps a bit :)
@1prakharchoudhary5 ай бұрын
Great video Cindy, I hope this video get viral, remember me after that <3
@TechCindy5 ай бұрын
Thanks, I'll remember you for sure! Have a nice lovely day
@ethanhe425 ай бұрын
very helpful for beginner! hope I knew 10 years ago
@TechCindy5 ай бұрын
Thanks! You must be very senior in the field. Feedback and suggestions all welcomed :)
@AtomicSlugg5 ай бұрын
great video! ❤ i've tried some workflows that involve tripo and autoremeshing and i completely agree, the result is just not worth the effort. a professional pipeline is so much more efficient and cleaner, with better control and iteration. i think there is potential in blender's geometry nodes combined with LLMs
@TechCindy5 ай бұрын
Thanks a lot for sharing!! I could tell you're a pro🔥 I also agree that there should be some new ways of doing it. Blender is so powerful and I look forward to better tools.