0:00:00.160,0:00:05.040 Greetings everyone. Today I am going  to introduce you to MimicPC. MimicPC   0:00:05.840,0:00:12.760 is a cloud service that lets you use AI  applications seamlessly in your browser,   0:00:12.760,0:00:19.200 while everything runs on remote cloud servers,  providing permanent storage as well. So,   0:00:19.200,0:00:25.560 I have prepared this post where you will find  relevant links and explanations. This is a   0:00:25.560,0:00:30.200 public post. The link of the post will be in the  description of the video. So, how are you going   0:00:30.200,0:00:37.360 to register? Open this link. Then you can either  register by email or using your Google account,   0:00:37.360,0:00:42.560 which I am going to use my Google account.  Continue with Google. I will use my email.   0:00:42.560,0:00:50.000 Continue. When you register first time, they will  give you 50 cents. You can set a password to login   0:00:50.000,0:00:54.440 also, if you wish. I am just going to skip it  because I have used the Google account. And on   0:00:54.440,0:01:00.160 this screen you will see that I have balance. I  can add balance from add funds from here. Also,   0:01:00.160,0:01:05.400 you will see that they are giving us  50 gigabytes storage here. You see,   0:01:05.400,0:01:11.680 this is a permanent storage. Whatever the file I  upload, my models, my generated images, my videos,   0:01:11.680,0:01:17.840 whatever, they will be permanently saved. And when  I am using next time, they will be there. Also,   0:01:17.840,0:01:24.040 I will be able to access my files without spending  any money, without starting any service. This   0:01:24.040,0:01:30.160 is a very, very good deal. They are giving 50  gigabytes. And you can also upgrade the storage   0:01:30.160,0:01:36.360 if you need. You see, you can upgrade it to 150  gigabytes, 250, and so on. This is a very cool   0:01:36.360,0:01:43.800 feature. So, this is the main screen. You can go  to the add new app and select the existing apps   0:01:43.800,0:01:50.800 from here. You see Automatic1111 web UI, Fooocus.  Then there is RVC. You can also search here,   0:01:50.800,0:01:55.520 like Whisper. Let's search it. Okay, looks like  the search is not working very well. There is   0:01:55.520,0:02:01.160 ComfyUI. You can install ComfyUI, load  your workflows and use them. You can use Face   0:02:01.160,0:02:09.640 Fusion, Chatbot, AI Toolkit, and ChatTTS and  others. They are keep adding new applications   0:02:09.640,0:02:15.240 to here. Moreover, they have a Discord server.  When you join their Discord server from this link,   0:02:15.240,0:02:21.240 you will get into this page. And in this page,  there are free workflows. So, you can use the   0:02:21.240,0:02:27.680 free workflows of other people. You can chat with  the developers. Ask your questions and anything.   0:02:27.680,0:02:33.080 This is very, very useful. So, let's continue.  I am going to show a few applications in this   0:02:33.080,0:02:40.560 tutorial. First of all, the CogVideoX-Fun. This  is the latest open source. The very best open   0:02:40.560,0:02:45.520 source image to video model. So, how you are  going to start it. There is a quick link here.   0:02:45.520,0:02:51.720 This is a shared workflow. You can also share  your workflows. I will click apply from here,   0:02:51.720,0:02:59.400 and I am logged into my account. Then it will ask  me to select a GPU. There are several tier GPUs,   0:02:59.400,0:03:05.080 as you are seeing right now. Medium, large, large  pro, ultra, and mini. Based on the workflow that   0:03:05.080,0:03:11.720 you are going to use. You need to select accurate  hardware. Since CogVideoX is extremely hardware   0:03:11.720,0:03:17.960 demanding. I am going to use ultra. Ultra  gives me L40s, 48 gigabyte GPU. This is a very,   0:03:17.960,0:03:24.800 very powerful GPU. Then you can set your machine stop after certain minutes. You can set automatic   0:03:24.800,0:03:29.960 extension. This is very useful if you forget to  turn off your machine, it will be automatically stopped.   0:03:29.960,0:03:34.600 You see, there is automatic stop as well. So,  let's click the create and start. It will start   0:03:34.600,0:03:41.600 the machine. This is going to auto load models  from their system since this is a shared workflow,   0:03:41.600,0:03:47.960 and in the right panel you see my files are here.  All of the files that I have used will be saved   0:03:47.960,0:03:53.200 here entirely, and I will be able to access  them. So, let's return back to our dashboard,   0:03:53.200,0:04:00.000 and this is the newest machine that, it is getting  started. So, we are waiting. I suggest you to   0:04:00.000,0:04:07.560 name them to not forget them. So, click this  icon and let's give CogVideoX fun model. So,   0:04:07.560,0:04:13.633 you should name them to remember them later. And  if you want to access their files, you see there   0:04:13.633,0:04:18.800 is go to manage, when you go to manage, you will  see all of your files here without starting them.   0:04:18.800,0:04:23.760 Let me show you. For example, in this machine,  I have generated some videos. Go to manage. This   0:04:23.760,0:04:29.320 machine is not started, so I am not using any of  my credits. And in the output. In the right panel,   0:04:29.320,0:04:35.160 you see the videos that I have generated. For  example, I can download this video and show you   0:04:35.160,0:04:41.520 that I have generated. Let's start it. You see,  this was animated video from a static image,   0:04:41.520,0:04:48.120 so I can access all of my files from here. I can  upload models according to my needs. You see,   0:04:48.120,0:04:54.320 there is upload icon here. So, go to the models  and click upload. You can also download models   0:04:54.320,0:05:00.760 from remote servers. So, when I click here. It  will allow me to choose either from my computer   0:05:00.760,0:05:07.480 or upload from URL. Like, you can upload  from HuggingFace repository, from CivitAI,   0:05:07.480,0:05:14.960 or similar FTP supporting servers with direct  URLs. And my machine has started. You see,   0:05:14.960,0:05:23.200 I click operate and this is the CogVideoX fan  model. This is using the ComfyUI as a backend. So,   0:05:23.200,0:05:29.160 you don't need to do anything actually. Okay.  Looks like this link has been changed to   0:05:29.160,0:05:35.440 something else. It says FLUX model, which is  not supposed to happen. See, I will fix this   0:05:35.440,0:05:41.720 link. Yes, this is CogVideoX. Maybe I started the  inaccurate one. Let's start this one. Apply. And   0:05:41.720,0:05:47.040 let's select the hardware as Ultra. Create and  start. Maybe I clicked the inaccurate link. I'm   0:05:47.040,0:05:54.360 not sure. Okay. Now it's starting. Return back to  our session. So, this is getting started. This was   0:05:54.360,0:06:00.480 FLUX. Don't worry. I will fix this link if it is not  the CogVideoX model. Okay. This time we got the   0:06:00.480,0:06:06.160 accurate model. So, also I am going to share this  another machine with you, and I am going to show   0:06:06.160,0:06:12.640 you how to share. You see, there is share option.  Let's click this, and I will say create image. So,   0:06:12.640,0:06:19.200 this way I will be able to share my image with  you, and you will be able to use it. Let's   0:06:19.200,0:06:26.360 name this as FLUX. Okay. You can also add image  description. I will just save. And then I will  0:06:26.360,0:06:32.920 have this link, and I am going to add it here, so  you can also access it quickly. Let's add it here.   0:06:32.920,0:06:38.920 So, now when you open this link that I have  shared, you will be able to use the workflow   0:06:38.920,0:06:45.760 that I have. And where was it? Let's go back to  our operate. It was here. Let's see. And yes,   0:06:45.760,0:06:51.880 for example, this is using FLUX model right  now. You can also use all other models that   0:06:51.880,0:06:57.760 are shared in their system. So, this is how their  system works. This is just amazing. For example,   0:06:57.760,0:07:04.600 let's use this FP8 model and generate an image,  and let's return back to our video model. So,   0:07:04.600,0:07:10.560 this is the CogVideoX model. Let's upload an  image. So, I'm going to animate this image as   0:07:10.560,0:07:18.040 an example, and how you need to prompt. This  is still in research. I am also not sure how   0:07:18.040,0:07:24.760 exactly to prompt. I have made some testing, and  I have prepared a quick prompt for you. You see,   0:07:24.760,0:07:32.760 you can use POE and Claude 3.5 or ChatGPT or  Gemini to generate a caption. I have written   0:07:32.760,0:07:39.120 a quick prompt here. So, let's write the prompt  here. This is a free version that is allowing you   0:07:39.120,0:07:47.000 to use, free. Upload your image from your computer  and get a caption. However, as I said, not all of   0:07:47.000,0:07:53.320 the descriptions that it generates works. So, this  is somewhat you need to get experience. I am still   0:07:53.320,0:07:59.800 not sure. So, let's copy this one. Let's return  back and paste your prompt here. That is all you   0:07:59.800,0:08:06.680 need to do. Then click Queue prompt. It will load the  model and generate a video from this. And when we   0:08:06.680,0:08:13.160 return back to our ComfyUI workflow with FLUX. It  generated the image. However, it is broken because   0:08:13.160,0:08:19.080 FLUX model doesn't work like this. You need to  set the CFG 1 and we need to set the resolution   0:08:19.080,0:08:30.240 to like 1024 to let's see 1024, 1024. Actually, I will  update my image so it can be like this. Then   0:08:30.240,0:08:37.480 as a sampler, I find that iPNDM is working  best. So, let's generate another image. Now it   0:08:37.480,0:08:42.600 should be accurate. You see, the speed is really,  really good. This is running on a cloud machine,   0:08:42.600,0:08:49.080 not on my computer, and image is generated.  Now I am going to share this. Actually see,   0:08:49.080,0:08:55.360 you do not have enough storage to share. Okay,  I need to expand my storage. So how I can expand   0:08:55.360,0:09:00.080 my storage. I'm going to delete some of the  applications that I have. So, when I delete them,   0:09:00.080,0:09:07.960 they will empty my storage. Delete, delete, and  it's also delete this one for now. Okay, I have   0:09:07.960,0:09:15.000 got more storage as I delete other ones. They  will be also fixed. Okay. And this was another   0:09:15.000,0:09:22.880 one started. So, let's see which ones are, which  ones so let's delete this one. Okay. It's okay.   0:09:22.880,0:09:28.200 Ok this is gone. Accurate image. Return back to  this one. This is also accurate one. It is working   0:09:28.200,0:09:35.360 fine. So, I want to share this workflow with you  as a FLUX development model. FP8. You can also use   0:09:35.360,0:09:42.640 this on less powerful hardwares, cheaper ones.  So, I am going to share this as an image. Yes,   0:09:42.640,0:09:50.960 not important. FLUX dev. Okay, save and share.  Okay, I can share indefinitely. Create sharing,   0:09:50.960,0:09:58.360 copy. That's it. And now once you click this new app,  it should work. So, this is how you can share   0:09:58.360,0:10:04.160 applications. You also don't need to use  mine, why? Because, as I said, official   0:10:04.160,0:10:09.800 applications already have them. For example,  let's also make example of Automatic1111 web UI,   0:10:09.800,0:10:16.160 and this can run pretty much on every hardware.  For example, let's start with this T4. Create   0:10:16.160,0:10:22.000 and start Auto1111 web UI. We should also  name it as Auto1111 web UI. Okay, actually   0:10:22.000,0:10:28.440 it shows that the name. Then let's create another  application as new application and Fooocus for   0:10:28.440,0:10:35.080 example. You can also use Whisper. Actually let's  also make a Whisper example. Get started. Create.   0:10:35.080,0:10:41.440 Whisper also runs on all of the hardware, and  let's also make another one as Fooocus. You see   0:10:41.440,0:10:47.680 the logic is same. You can also upload your  models, your configurations, everything works,   0:10:47.680,0:10:53.000 everything is saved in your permanent storage.  They give you 50 gigabytes of permanent storage   0:10:53.000,0:11:00.920 every month, as I said, for free. And this was our  FLUX model. Until you stop your running instance,   0:11:00.920,0:11:06.520 it will use your credits, but once you stop it,  it will give back your credits. So, when you   0:11:06.520,0:11:12.320 first time started, it will deduct your credits  from your account, but when you stop it back, it   0:11:12.320,0:11:17.840 will return back the remaining time. So, what does  this mean? Let me show you. Let's refresh. I have   0:11:17.840,0:11:24.640 currently around 9 dollars. So, when I stopped  this one, it should return back to my credits. So,   0:11:24.640,0:11:31.360 you see, my credits are returned back. This is how  their system works. And on this page, you see home,   0:11:31.360,0:11:38.120 mini PC dashboard, it shows which instances  are running right now. So, I have CogVideoX,   0:11:38.120,0:11:44.600 Automatic1111 web UI, Whisper web UI, and Fooocus.  Let's close all of them because we can reconnect   0:11:44.600,0:11:50.160 them. Actually, this is working. So, let's just  keep it as it is. Close all other ones. And for   0:11:50.160,0:11:55.360 example, let's go to the this one. Auto1111 Web UI. Starting. So it is started. 0:11:55.360,0:12:01.197 So in here I can use all these models. You see they are all added here. This is just amazing. 0:12:01.197,0:12:03.993 You can select any model and use directly. 0:12:03.993,0:12:09.621 For example lets select one of the models from here. For example ColorFull XL V6. 0:12:09.621,0:12:18.200 You can also use your own models, CivitAI models Hugging Face models, how you can use? 0:12:18.200,0:12:22.361 I am going to show example from Hugging Face. For example RealVis XL V5 0:12:22.361,0:12:29.200 and you want to use this model, go to the  RealVis XL version 5 fp16. Copy download link   0:12:29.200,0:12:35.080 and go back here and in the models click here,  and this goes into the Stable Diffusion folder,   0:12:35.080,0:12:40.800 as you know it. Click upload. I am going to upload  it from remote server like this. So, you see, this   0:12:40.800,0:12:47.880 is the download URL. Click upload and just wait,  it will upload it. Unfortunately, it doesn't show   0:12:47.880,0:12:54.400 the status, so you need to wait the upload to be  completed. I think it is going to show in the logs   0:12:54.400,0:13:00.200 or somewhere. Let's see, it doesn't show here. It  was showing somewhere. Okay, I think you need to   0:13:00.200,0:13:07.480 click here and just wait until it is completed.  So, let's go back to models and close this   0:13:07.480,0:13:13.240 upload icon. Let's go back to Stable Diffusion  and refresh. So, you just need to wait until it is   0:13:13.240,0:13:19.080 completed. I told this to the developers. I think  they will add it. Currently SDXL Colorful XL is   0:13:19.080,0:13:26.920 loaded, a car. Okay, let's go to a simple thing,  a car. Let's select one of the best sampler, UniPC,   0:13:26.920,0:13:34.640 and 1024 like this. Let's generate an image. This  is SDXL model. It shows that SDXL, and our video   0:13:34.640,0:13:41.600 is still getting generated. Unfortunately, this  video model extremely demanding and slow, but I   0:13:41.600,0:13:47.520 think there will be better models soon coming  up. So, you can also use them. And I was also   0:13:47.520,0:13:53.280 going to show you LivePortrait video to video for  example. Let's go to the video to video. Let's   0:13:53.280,0:13:59.240 open this image, and also there is image to video LivePortrait. Let's also open them.   0:13:59.240,0:14:06.760 You just need to click apply. You can use a lesser  hardware here, like large pro. Create. And here   0:14:06.760,0:14:14.160 apply. Okay. And let's also do large pro. See, it  started. Yes, here. Since I am starting multiple   0:14:14.160,0:14:19.760 things, it may be confusing but when you start one by  one you won't get confused. It is just so easy,   0:14:19.760,0:14:25.040 and it is getting started. And meanwhile, let's  look at the Auto111 web UI. Okay. This   0:14:25.040,0:14:30.680 image is generated with this model and our model  arrived here. You see, it was downloaded from the   0:14:30.680,0:14:36.680 HuggingFace. So, when I refresh here, I should be  able to see it. It is getting, refreshed. Yes. Let's   0:14:36.680,0:14:43.680 type its name and you see RealVis XL arrived.  This is permanently saved in my folders right now.   0:14:43.680,0:14:50.480 So, it is getting loaded. Just wait until the  load completed. This is the same as in Fooocus,   0:14:50.480,0:14:55.480 in ComfyUI. You see, I can connect to the  Fooocus, and I can just start using it.   0:14:55.480,0:15:00.280 And there is also a Whisper web UI. I can start  using it. For example, this is the Fooocus. Let's   0:15:00.280,0:15:07.240 use the advanced. And in here, models, all of  the models are also here. However, I also need to   0:15:07.240,0:15:13.160 download the model that I want into here as well.  So, currently it is SDXL base, like car. It will   0:15:13.160,0:15:19.440 just start generating. Fooocus. And there is also a  Whisper web UI, as you are seeing right now. Just   0:15:19.440,0:15:25.200 select an example video, upload it, and generate  subtitle. It will automatically extract the audio   0:15:25.200,0:15:31.000 and generate subtitle. This is Whisper web UI.  I'm not going to delve into how to use these   0:15:31.000,0:15:37.280 applications. They are separate applications, but  I am just showing you how you can use Mimic PC.   0:15:37.280,0:15:42.600 Everything is running on remote server, but I am  using as it is on my PC. You see, and it doesn't   0:15:42.600,0:15:48.800 install anything. Because this is running on my  browser, nothing is installed into my browser or   0:15:48.800,0:15:54.560 on my computer. All of this is running on a remote  cloud server. And the images are getting generated   0:15:54.560,0:16:03.120 in the Fooocus web UI. Okay. So, this is loaded  screen of the LivePortrait. For example, this is   0:16:03.120,0:16:09.400 from image to video. Okay. Let's see. Yes. I have  demo materials on my computer. So, let's say this   0:16:09.400,0:16:15.080 is the image. Let's load a video like this. This  is the latest version of the LivePortrait. So, you   0:16:15.080,0:16:23.320 can change the like, ee, woo, and other things of  the image, like here. Let's make it more smiling.   0:16:23.320,0:16:29.200 Okay. There is expression editor. Then just let's  queue the prompt and let's wait it. And this is   0:16:29.200,0:16:35.400 another LivePortrait. This is only image. Okay.  So, you can play with this and see how they are   0:16:35.400,0:16:41.440 working. Let's queue prompt. And the Whispers also  generated the transcription. It's pretty fast.   0:16:41.440,0:16:47.320 Okay. This file, this audio, I think. Yes. This  is one. Okay. Let's generate subtitle. It should   0:16:47.320,0:16:53.200 be very fast after first time. Yes. You see, it  is almost instant. So, it generated the subtitle.   0:16:53.200,0:16:59.000 You can also download the subtitle from here. Just  open it and you see, this is the subtitle. Okay.   0:16:59.000,0:17:05.720 Everything is working. This also generated a GIF  image from the uploaded image. You see, like this.   0:17:05.720,0:17:12.120 So, it is like blinking its eye. You see, you can  also set another things here. For example, let's   0:17:12.120,0:17:17.000 set 10 smile or 1.3. Let's  queue prompt. This should be pretty fast. Yes.   0:17:17.000,0:17:23.120 Now it is also smiling. Okay. Yeah. Pretty good.  You see. Both smiling and blinking. And here the   0:17:23.120,0:17:30.840 original image. And this was expression transfer  from video. Okay. Currently it doesn't save the   0:17:30.840,0:17:36.920 audio. So, you also need to save the audio. How  you gonna save it? To save the audio, you just   0:17:36.920,0:17:42.240 drag and drop the audio and requeue. By the way,  everything will be in the output folder. So, you   0:17:42.240,0:17:49.040 can also download them later. Okay. It is getting  generated. And let's look at our image animation.   0:17:49.040,0:17:56.080 Yes. You see, this static image, turned it into  this video. Okay. Yeah. It's a very weird video.   0:17:56.080,0:18:03.040 As I said, how you prompt is a mystery right  now. People are working on them. But my prompt   0:18:03.040,0:18:10.480 generation also works decently. It really animated  the image into this video. This is a free video.   0:18:10.480,0:18:15.760 You don't wait. You just run it on Mimic PC. You  can also run it on your computer if your computer   0:18:15.760,0:18:21.680 is powerful. And yes. This is the newer video  getting generated. Yes. Currently it is combining   0:18:21.680,0:18:29.120 the video. You can also follow the status on  this green line here. Okay. And video generated.   0:18:29.120,0:18:35.440 Refresh. Refresh here to get the video. Yes. The  video arrived. I can click here and download.   0:18:35.440,0:18:41.880 Let's play it. It's fresh. And then the other big  thing for me is sleeping. I love sleeping. And I   0:18:41.880,0:18:48.560 need. Okay. As you see, it generated video. You  need to play with settings to get better results.   0:18:48.560,0:18:55.120 And this is how you use Mimic PC. I hope you  have enjoyed. Please ask any questions. Please   0:18:55.120,0:19:01.160 join their Discord server and chat with the  developers and the experts there. They keep   0:19:01.160,0:19:08.800 adding new features, new amazing stuff. So, you  can use all of the applications and also your own   0:19:08.800,0:19:14.080 workflows with ComfyUI. It also has a default  ComfyUI. So, go to ComfyUI. Go to ComfyUI from   0:19:14.080,0:19:20.320 here and get started. This will start a fresh  ComfyUI for you. Let's start it to show. Then   0:19:20.320,0:19:26.240 you can load your configurations. It comes with  so many pre-installed features. For example,   0:19:26.240,0:19:32.040 in the Automatic1111 web UI, there are  ControlNet, After Detailer, you see Cutoff,   0:19:32.040,0:19:37.440 LLuL, I don't even know this. And there are some  other things like Segment Anything. So, it is   0:19:37.440,0:19:45.640 not a default one. It comes with extra features,  like Deforum. Photopea, OpenPose editor, Infinite    0:19:45.640,0:19:52.440 Zoom. There is also extensions tab here, which  you can use, tagger. So, it is not a default. It   0:19:52.440,0:19:58.880 is a more advanced version with extra features.  Let's look at the ComfyUI as well. A default   0:19:58.880,0:20:03.920 ComfyUI. It is getting started right now. And  this is the default ComfyUI. You can select   0:20:03.920,0:20:11.400 all the models as well. So, these are all shared  models. It also has ComfyUI manager installed,   0:20:11.400,0:20:18.280 as you are seeing. So, you can  load your configs and use them as it is in your   0:20:18.280,0:20:25.200 PC. However, this is running on a remote server.  So, it is working great. And that's it. I hope you   0:20:25.200,0:20:31.080 have enjoyed. Hopefully see you in another amazing  tutorial. I am working on fine tuning of FLUX. I   0:20:31.080,0:20:37.480 am doing a lot of testing right now. Testing batch  size 7. I am also running different testings,   0:20:37.480,0:20:44.680 experimentation and trainings to further improve my SDXL fine tuning. Hopefully see you later.