{"id":28545,"date":"2025-02-18T09:36:01","date_gmt":"2025-02-18T01:36:01","guid":{"rendered":"https:\/\/www.1ai.net\/?p=28545"},"modified":"2025-02-11T21:46:52","modified_gmt":"2025-02-11T13:46:52","slug":"stable-diffusion%e6%80%8e%e4%b9%88%e7%94%a8%ef%bc%9fstable-diffusion-lora%e6%a8%a1%e5%9e%8b%e6%80%8e%e4%b9%88%e7%94%a8%ef%bc%8clora%e6%a8%a1%e5%9e%8b%e4%bd%bf%e7%94%a8%e6%94%bb%e7%95%a5%e6%95%99-2","status":"publish","type":"post","link":"https:\/\/www.1ai.net\/en\/28545.html","title":{"rendered":"Stable Diffusion how to use? stable diffusion lora model how to use, lora model use strategy tutorial (below)"},"content":{"rendered":"

In the previous section, we learned the basics of Lora, including reading author and comment section information, understanding the model, and how to open the additional web interface and fill in the appropriate weight ranges.<\/p>\n

Today, we're going to move into a more in-depth study, including how to use trigger words, reference author parameters for generation, and make adjustments based on the results.<\/p>\n

I. Detailed steps for using Lora<\/p>\n

Step 1: Use Trigger Words<\/p>\n

Trigger words are key when invoking Lora. Sometimes, just calling Lora may not produce the desired effect, and positive cues need to be added to trigger it correctly. For example, when generating images, in addition to using weights and big models, specific cues such as \"full body\" and \"chibi\" need to be added to trigger specific styles.<\/p>\n

\"f685d278j00sritsj007gd000u000esm\"<\/p>\n

Step 2: Refer to the author's parameters for generation<\/p>\n

In the process of mimicking the author-generated process, we need to understand the macromodel, trigger words, and refer to parameters such as positive and negative cues used by the author, CFG, seeding, and sampling methods. With the consistency of these parameters, we can replicate almost identical images.<\/p>\n

\"e992a36dj00sritsj00a7d000u000esm\"<\/p>\n

Step 3: Adjustment based on generated results<\/p>\n

The final step is to adjust the generated results in many ways, including models, weights and various parameters to achieve the results we want.<\/p>\n

\"bd874995j00sritsj005nd000u000esm\"<\/p>\n

\"11815cc0j00sritsj006ad000u000esm\"<\/p>\n

II. Notes on the use of Lora<\/p>\n

There are a few things to keep in mind when using Lora:<\/p>\n

    \n
  1. Model name consistency<\/strong>: Ensure that the Lora name used is the same as the locally saved name, otherwise it may not be called correctly.<\/li>\n
  2. Beginner's caution with multiple Lora<\/strong>: Using multiple Lora requires some understanding of the underlying principles of Lora, and newcomers are advised to start with one Lora.<\/li>\n
  3. Generate images as cue charts<\/strong>: The generated image can be set as a preview image of Lora to help better understand what Lora does.<\/li>\n<\/ol>\n

    \"d746b69dj00sritsg003md000u0008mm\"<\/p>\n

    \"9e3c4aadj00sritsj004zd000u000esm\"<\/p>\n

    III. Practical case sharing<\/p>\n

    Next, we will demonstrate the usefulness of Lora with a few real-world examples.<\/p>\n

    1\u3001Picture details in high definition<\/p>\n

    By using a specific Lora, we can increase the detail and enhance the texture of the image. For example, when generating a still life image of a product, adding a detail-enhancing Lora can make the image look more realistic.<\/p>\n

    \"fc656256j00sritsk00c5d000u000esm\"<\/p>\n

    \"77eddfcaj00sritsw00zdd000u000k0m\"<\/p>\n

    2. Character traits are fixed<\/p>\n

    In scenes where character features need to be consistent, a specific Lora can be used to fix a character's facial or clothing features.<\/p>\n

    \"6a0ef986j00sritsl00e4d000u000esm\"<\/p>\n

    \"98c1ae94j00sritt701upd000u00140m\"<\/p>\n

    3. Characters de-AI'd<\/p>\n

    By using a specific Lora, we can reduce the unreality of AI-generated characters and make the images closer to real photos.<\/p>\n

    \"2ad899dej00sritsk009vd000u000esm\"<\/p>\n

    \"0c565f5ej00sritt9026hd000sg01a8m\"<\/p>\n

    4\u3001Line drawing generation<\/p>\n

    For designers and artisans, Lora can be used to generate line art effect images, which are useful for creating illustrations and design elements.<\/p>\n

    \"fd070f3cj00sritsk008sd000u000esm\"<\/p>\n

    \"120e20c3j00sritt602qhd000sg01a8m\"<\/p>\n

    5\u3001Picture stylization<\/p>\n

    Lora can also be used to add a specific style to an image, such as Makoto Shinkai style, and by learning a specific animation style, you can give the resulting image a specific artistic effect.<\/p>\n

    \"cfb7d73ej00sritsj00azd000u000esm\"<\/p>\n

    \"2c5a17ccj00sritsw015hd000u000k0m\"<\/p>\n

    IV. Conclusion<\/p>\n

    I hope this article will help you understand and use Lora better.<\/p>\n

    We have not only refined how to use Lora, but also shared some practical examples. This knowledge will provide a solid foundation for our future project cases.<\/p>","protected":false},"excerpt":{"rendered":"

    In the previous section, we have learned the basics of Lora, including reading author and comment area information, learning about models and how to open additional web interfaces and fill out appropriate weight ranges. Today, we are going to go into more in-depth learning, including how to use trigger words, reference author parameters to generate, and to adjust to the results. First step: the use of the trigger word is the key to the call. Sometimes the mere call to Lora may not have the desired effect and a positive hint is needed to be correctly triggered. For example, in generating pictures, in addition to the use of weights and large models, specific hints, such as \u201cfull body\u201d and \u201cchibi\u201d, are needed to trigger a particular style. Step 2:<\/p>","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[149,144],"tags":[2328,197,198],"collection":[262],"class_list":{"0":"post-28545","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"hentry","6":"category-jiaocheng","7":"category-baike","8":"tag-ai","9":"tag-stable-diffusion","11":"collection-stablediffusion"},"acf":[],"_links":{"self":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/28545","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/comments?post=28545"}],"version-history":[{"count":0,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/posts\/28545\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/media?parent=28545"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/categories?post=28545"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/tags?post=28545"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/www.1ai.net\/en\/wp-json\/wp\/v2\/collection?post=28545"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}