AI视频的一下相关技术
Fooocus可以生成逼真的图片。
Jaggernaut V9 基本模型the base model
拥有GPU的电脑,或者借助平台类似: platforms like RunDiffusion
要做到让模型动起来,除了借助技术外,还需要很多东西:
Not just that. Extra content, free stuff, exclusive photos, lingerie shots, topless and nudes.
But I'm using Juggernaut V9 mostly as base model. 基础的模型是用 Juggernaut V9, 一致的图像的问题,提示词工程是必须的。。。
For consistent images, prompt engineering is a must.
Hey. I’m working on Fooocus UI with different SDXL models locally. 我们使用 Fooocus UI 里的不同的SDXL模型。 我们有一个提示词工程的工作流。 I can’t reveal so much about consistency - but we have a workflow with prompt engineering.
Jaggernaut V9 基本模型the base model
拥有GPU的电脑,或者借助平台类似: platforms like RunDiffusion
要做到让模型动起来,除了借助技术外,还需要很多东西:
Not just that. Extra content, free stuff, exclusive photos, lingerie shots, topless and nudes.
But I'm using Juggernaut V9 mostly as base model. 基础的模型是用 Juggernaut V9, 一致的图像的问题,提示词工程是必须的。。。
For consistent images, prompt engineering is a must.
Hey. I’m working on Fooocus UI with different SDXL models locally. 我们使用 Fooocus UI 里的不同的SDXL模型。 我们有一个提示词工程的工作流。 I can’t reveal so much about consistency - but we have a workflow with prompt engineering.
脚本和场景:ChatGPT
角色和背景图:Midjourney
背景移除:photoshop/ remove.bg
音效:Eleven Labs beta.elevenlabs.io
动画:RunDiffusion/StableDiffusion/Genmo(rundiffusion.com/stability.ai/genmo.ai)
Stable Diffusion:AI 图像生成工具,一个基于 Latent Diffusion Models(潜在扩散模型,LDMs)的文图生成(text-to-image)模型,生成图片的原理是通过模拟稳定扩散过程并生成相应的数据,然后将数据可视化展示为图片,以帮助用户更好地理解实验结果和数据特征
角色和背景图:Midjourney
背景移除:photoshop/ remove.bg
音效:Eleven Labs beta.elevenlabs.io
动画:RunDiffusion/StableDiffusion/Genmo(rundiffusion.com/stability.ai/genmo.ai)
Stable Diffusion:AI 图像生成工具,一个基于 Latent Diffusion Models(潜在扩散模型,LDMs)的文图生成(text-to-image)模型,生成图片的原理是通过模拟稳定扩散过程并生成相应的数据,然后将数据可视化展示为图片,以帮助用户更好地理解实验结果和数据特征
图像领域大火的深度生成模型Diffusion Model, 通过文本生成图像,文本和图像的学习CLIP模型,使用CLIP模型生成输入文字embedding
扩散模型Diffusion: Diffusion 模型的训练可以分为两个部分:
扩散模型Diffusion: Diffusion 模型的训练可以分为两个部分:
1. 前向扩散过程(Forward Diffusion Process)→图片中添加噪声;
2. 反向扩散过程(Reverse Diffusion Process)→去除图片中的噪声
2. 反向扩散过程(Reverse Diffusion Process)→去除图片中的噪声
Not really seeing this on here:
- Read up on Ollama or Stable Diffusion
- Have a decent powered GPU
- Play around with modeling to get good at it
- Create a separate IG account that labels it as AI
- Try your luck.
Modeling has gotten really good now with little effort needed
if you are wondering how to make a fully generated AI video
look no further I will show you exactly what tools I used and how you can get the same results like
mine which can be seen at the end of this video let's start with first things first for a good
video you need a script characters scenes moving heads and some voices right so here's the list
of tools that I used and let me show exactly how I used each one of them and then I will
show you how they work first one is chatgpt for ideas scripts conversations and scene write-ups
then we used MidJourney for characters and scene designs remove.bg for background removal which is
behind our characters Eleven labs for artificial voices that are the most similar to my characters
and then a combination of stable diffusion run diffusion and genmo for animation of
the images the ID for Talking Heads and vision to generate tweets but this can be done through many
different tools as well as the code inspect and lastly cartoon face as I'm still not well known
I had to recreate my head with a simpler tool because MidJourney was giving me okayish results
but I really wanted something better so how did I actually use these tools from ChatGPT I needed
the ideas and what the actual video is going to be about so I first fed the tool with freelancer.com
contest that I wanted to participate in and added a prompt for it to create potential directions
for my new video once I got the creative direction for my video I instructed a tool to create new and
different scenes conversations and characters as well as some celebrity cameos and lastly I needed
the name for my show so I prompted chatgpt to give me a number of potential names this is more of a
reference for the video to get a great name you need to play around a bit with it and provide
feedback to look at the desired results but it shouldn't take long thanks to the airprm tool for
chatgpt you can open a new thread or with airprm find the MidJourney prompt generator and then
just copy and paste the adjusted brief and scenes from chatgpt which input straight to Midjourney
next on my list was Midourney once I actually got my script scenes and characters I needed
to find real life images of those mentioned characters because Midjourney will not create
a good looking image of The mentioned celebrity without the reference image I will show you just
an example and you will need to replicate this setup for all of the characters in your video
for those of you that are new here Midjourney is a brilliant tool that can produce amazing images
just out of plain text and with every prompt in Midourney you start with first adding /imagine
then you go and copy the image URL of the desired person and here for us we have Margot Robbie and
then you paste that link into the prompt once that is done you can use the prompts from airprm which
is a chatgpt plugin for prompt creation or create them yourself in my case because my show is set
on the island I will have a very descriptive words and lastly and most importantly I need
my style and my parameters I did a very specific style for its reference and an AR of 16:9 which
is a resolution because I will need this format later for my YouTube video make sure to copy it
like I did because just a small mistake and it won't work also I will need my background scenes
so I just reuse the previous prompt to create consistency without the URL of Margo next thing
that we need to do is to remove background from our characters and add the scenes in behind and
this part is super easy I used Photoshop because I know how to use it and it's a bit easier for me
but you can go to remove.bg if you do not have experience add the AI generated photo of Margo
wait for the magic to happen and there you have it now you can use her image to paste it on all other
backgrounds that you have created such as the one we created from the same prompt next one is a
brilliant AI which is called Eleven Labs there is a big disclaimer here I used the voice of Margo's
solely for educational purposes but that is her own property and it shouldn't be used you need to
be sure that you have the right to use anyone's image or anyone's voice and have their consent
now that we got that out of the way in Eleven Labs we go on advice then instant voice cloning we name
it describe it and give it a few prompts that so the tool can understand how it should sound
like and then generate once we click generate we have our new voice we click use and you will see
that you have a couple of voice settings like speed Clarity or how similar or different it
is going to be and lastly and most importantly in the text box the add the text which actually
needs to be created for image animation I started with stable diffusion first and its own online
variation called Rundifusion But realized that I will need too much time to do it so I found a much
simpler solution called Genmo I will soon create a separate tutorials on both stable diffusion
and run diffusion but for now let's stick to Genmo Genmo is actually short for Generation
plus motion and it's a tool that generates video from text with the help of this tool
we generated all the movements in the scenes we generated using Midjourney in our previous steps
there are three options in general for creating the first frame or set search create and upload
we used upload where we uploaded the scenes that were previously generated we also need
to adjust the options how long do we want the video to be how drastic changes do we want how
fast should the scene change and how smooth do we want the transition between the frames some of
our guidelines were that we didn't want to change the scene too much so we kept exploration around
30 and dynamism around 17 to 20. it takes about three minutes for the tool to generate the entire
video for us so that we can decide if it's good enough or do if we want to generate it again with
different parameters now that we have our scenes characters and voices we actually go in the ID
and give it a life we go on to create video and to add our new presenter and here you can do it
directly with the background that you created earlier then you upload a voice audio upload
your own voice and once that is done you simply generate your video how brilliant and easy is that
and lastly a couple of very easy generators tweetgen to create feeds
that are similar or same like mine Star Wars generators to create similar looking
text but different fonts and music because of the IP of course and then face cartoon
or similar apps to create faces out of images as cartoons that's only in case
that midjourney is providing you with correct Solutions so let me show you our final result
here and if this was helpful like share subscribe and see you in the next one
阅读量: 618
发布于:
修改于:
发布于:
修改于: