HuangTaobox

This is a Guide For⭐️⭐️PRO⭐️⭐️Anime Come True这是一篇⭐️⭐️PRO⭐️⭐️Anime Come True的使用指南Although the interface of Tensor is already quite user-friendly, it's still necessary to explain how to use model combinations to achieve the best results.Part One:Regarding the selection of second-generation models, as shown in the image below,The red box is suitable for Asian facial features, while the green box is more appropriate for Caucasian facial features.虽然tensor的界面已经很简单了，但还是需要为大家讲解一下模型组合的使用，从而达到一个最佳的效果第一部分：关于second generation的模型选择，如下图所示，红色框适合亚洲人面孔，绿色框适合欧美人面孔Part Two:Regarding the Real level, it's generally recommended to use 'low' for a more accurate restoration of ACG images. If you notice a significant number of errors in the limbs, it suggests a conflict between the information provided by the ACG images and the realistic style. In such cases, you may opt for 'medium' or 'high'.第二部分：关于Real level，通常情况下用low，可以更好的还原ACG图片；如果发现肢体错误比较多，说明ACG图片提供的信息与写实风格有冲突，就可以选择medium或者highPart Three:Regarding the selection of the first-generation model,The red box is more suitable for realistic illustrations, while the green box is better suited for anime with a higher ACG content.第三部分：关于first generation模型选择红色框适合比较写实的插画，绿色框适合ACG含量较高的animeIf you want better limbs, go to first generation and choose the [photorealism pony v1] model如果你想要更好的肢体，那就first generation中选择[photorealism pony v1]模型Example:Anime with a high ACG content + Asian facial features.ACG含量较高的anime + 亚洲人Realistic illustration + Caucasian facial features.That concludes the basic tutorial. We wish you all a delightful experience!比较写实的插画+欧美人以上就是基本的教程，祝大家玩的愉快！

This time I want to train an image of Optimus Prime from the transformers one movie这次我想训练一个 transformers one电影中擎天柱的形象The first step is to create a training set of images. You can get a lot of materials by searching on the Pinterest website.首先是图片训练集的制作，到Pinterest网站去搜索，可以得到大量素材Go to the tensor website, select online train, and then upload the image进入tensor网站，选择online train ，然后upload imageSelect the image to edit tags and remove the tags that are most relevant to the character in the image. This is counterintuitive, but it must be done选择图片编辑tag，去掉图片中和角色最相关的tag，这点事反直觉的，但必须如此Select the model. I chose DreamShaper 2.5D here.选择模型，我在这里选择的是2.5d的dreamshaperWrite a trigger word. It is best to create a combination of your own, such as "rb123”写一个Trigger words，这里最好是自己创作的组合，例如“rb123”Write Model Effect Preview Prompt, which can be Trigger words➕Simple action写上Model Effect Preview Prompt，可以是Trigger words➕简单动作Click start train below to start training!点击下面的 start train就可开始训练啦！The training progress can be viewed in the training section of the personal center.训练进度可以在个人中心的training中查看The default training process is 10 rounds. In each round, there will be 4 pictures previewed. The top has the loss value of each epoch. Generally, the lowest loss value performs best.默认的训练过程是10轮，在每一轮会有4张图片预览，最上方有每个epoch的loss值，一般情况下最低的loss值表现是最好的选择你认为最好最相似的一个epoch，点击发布，就拥有你自己的lora模型啦！谢谢你的阅读，有不明白的部分，欢迎留言询问Choose the epoch you think is the best and most similar, click publish, and you will have your own lora model! Thank you for reading. If you have any questions, please leave a message.

Often, we have a picture in our mind and can find a similar picture, but we don’t know how to write a prompt. Although tensor provides a reverse inference tool, there are other processes such as copying and pasting in the middle, and it does not support NSFW.In short, filling in the options is a very troublesome thing. I am not the only one who thinks so!Using the workflow to make a small tool can simplify a lot of processes. You can see the various small tools I made. Basically, there is no need to write prompt words, because I am a very lazy artistThe following is a simple tutorial to teach you how to make the first AI Tool. It is very simple. Just follow my picture step by step!Step 1: Create a new workflowStep 2: Select the img2img templateStep 3: Double-click the mouse on a blank area of ​​the interface, search for [wd] in the interface that appears, and select the [WD14 Tigger] plug-inStep 4: Drag the image on the load image panel and connect it to the image on the WD. This is the basis of the workflow [connecting nodes]!Step 5: Change the WD14 modle to V3 version, which is the latest image reverse model. With it, you can change your image to prompt.Step 6. Right-click on the Clip Text Encode panel and select Covert Text to Input.Step 7: Double-click the blank area again and enter string functionStep 8: Right-click on the string function panel and click [convert text_b to input]; then connect [string] on the WD14 panel to [text_b] of [stringfunction]Step 9: Connect the string in the stringfunction panel to the text in the CLIP Text encode panel, so your image will become a positive prompt!Step 10: Are you tired of reading? I am also tired of writing, let’s take a break😀😀😀😀😀😀😀😀Step 11: Click the load checkpoint panel ckpt_name, you can select the model, this time we choose a pony modelStep 12: In the string function and another clip text encode panel, fill in the pony's mass prompt[positive]:score_9,score_8_up,score_7_up, [negative]:score_3,score_2,score_1Step 13: It's almost done! Set the numbers in the ksample panel, refer to my values:Step 14: Click upload in the load image panel, select the image you like (the longest side should not exceed 1280), then click generate, and that's it!Step 15: Click Publish in the upper right corner, and then select Share Workflow. You will have your own workflow tool. You can find and run it on your personal homepage.This tutorial ends here. In the next issue, we will teach you how to convert WORKFLOW into a gadget and make it more useful and complete. Thank you for your support! This tutorial ends here. In the next issue, we will teach you how to convert WORKFLOW into a gadget and make it more useful and complete. Thank you for your support!