Fiora LOL Preview

Good news, 3 updates of works have been completed this week :)

Bad news, pixiv’s trumpet account has also been blocked :(

This work made a new attempt. Because the effect of the character lora tends to be two-dimensional, the results of several large models were not very good, so in the end, one large model was used as the main image, and another large model replaced the face.

What is the final effect? I hope everyone can provide valuable suggestions!

Then there is the quality of recent works. You can also rate your favorite works. Your feedback will also be used to adjust the direction of future works!

Finally, I would like to thank all the fans who still support my works. The future creation of works is inseparable from your support and encouragement. Thank you all!

Another: Let me share with you some of the factors I am currently considering when creating my works.

1. The number of works. Currently, there will be two stable updates every week. There are no plans to adjust it for the time being. The main reason is that there are currently a large number of works in each issue. The number of pictures in the three works together has reached 400. And at present, my habit is to photoshop every picture to facilitate correction of AI logic errors.

2. The content of future works will mainly focus on fans’ comments and private message requests. However, in view of the current slowdown in the number of subscribers, it may be adjusted to $20 per month for subscribers to customize and create content.

3. Regarding the current tiers of 6 US dollars, 8 US dollars, and 10 US dollars, fans can make suggestions on whether the current price is reasonable (you can compare the charging standards of other creators)

4. Let’s talk about some details of the creative work:

     In the early days of creating my works, the AI generation effect was relatively poor. Some poses could only be created through openpose, so the number of works at the beginning was small, but the effect of the works was still very good. The disadvantage was that each work had a fixed pose. same.

     After more than a year, the current AI generation effect has been much better, so the main postures of my works can also be controlled using parameter words, such as standing, sitting, looking back, lying down, squatting, kneeling, etc. Correspondingly, the problem of homogenization of postures in each work has also been improved. At the same time, the saved repetitive labor time can increase the quantity of each work and create more high-quality works.

     Recently, I have re-enabled openpose in my works to reproduce the classic postures of some characters in the works. The presentation of these postures requires me to spend a certain amount of time and energy, but I think it is still worth it. However, such pictures of each work There won’t be too many. I will try my best to ensure the quality of these pictures.

The above is all I can think of for the time being. If I have new ideas, I will update them on the homepage to share them with you.

好消息，本周完成了3次作品的更新:)

坏消息，pixiv的小号也被封掉了:(

这次的作品进行了新的尝试，因为角色lora效果趋于二次元化，试了几个大模型效果都不太好，所以最后采用了1个大模型为主图，另一个大模型替换脸

最终的效果怎么样，还希望大家能够提出宝贵建议！

然后就是近期的作品质量，大家也可以给自己喜欢的作品打下分，未来作品的调整方向也会参考大家的反馈意见！

最后，再次感谢依旧支持我作品的各位fans，未来的作品创作之路离不开你们的支持与鼓励，谢谢大家！

另：关于我创作作品目前考虑的几点要素和大家分享下

1、作品的数量，目前每周稳定会更新两期，暂时没有调整的计划，主要是目前每期作品的数量较多，3档作品加一起的图片数量达到400张了。且目前我的习惯还是每张图都要ps，便于修正AI的逻辑错误。

2、未来作品内容，主要还是围绕fans们的留言和私信要求进行。但鉴于目前订阅人数增长趋缓，可能会调整为每月20美元的订阅者，才可以进行作品内容定制创作。

3、针对目前6美元、8美元、10美元的档位，各位粉丝可以提出建议，现阶段的价位是否合理（可以对比下其他创作者的收费标准）

4、再说说创作作品的一些细节：

    我在创作作品的初期，AI的生成效果比较差，一些姿势只能通过openpose来创作，所以一开始的作品数量较少，但作品呈现的效果还是很好的，缺点是每个作品的固定姿势相同。

    经过1年多的时间，目前AI的生成效果已经好了很多，所以我作品的主要姿势也可以采用参数词进行控制，比如站立，坐下，回眸，躺下，蹲下，跪下等。相对应的，每个作品的姿势同质化问题也有所改善。同时节省下来的重复劳动时间可以提高每个作品的数量，也可以创作出更多优质的作品

    近期我的作品，又重新启用了openpose，来复刻一些作品角色的经典姿势，这些姿势的呈现需要我花费一定的时间和精力，但我觉得还是比较值得的，但每个作品的这类图片数量不会太多。我会尽量保证这些图的品质。

以上，暂时就想到这么多了，我有新的想法会在首页更新分享给大家的。