I have been passionate about generating images with AI for a long time now, trying to create beautiful images without making them too obscene. I seek beauty and creativity. I have tried to create models and LoRAs with my outdated hardware. Unable to afford powerful hardware or subscribe to CivitAi (due to cost reasons), I have tried to use the buzz from CivitAi to continue creating new models. Many people don't know how long it takes to create a good model, but it takes days or weeks. Recently, CivitAi decided to split the buzz into two (blue and yellow) and create the 'new creator program,' but it is only for subscribed users. I am not interested in making real money, but due to these implementations, many unsubscribed creators have been forced to leave. It is no longer possible to donate buzz if you don't have yellow buzz, and it is not possible to do so with blue buzz. Therefore, many users cannot donate anything. Only those who are subscribed or have yellow buzz can do it. I have also noticed that the 'Early Access' system does not work as it should. Many users neither download nor donate buzz and simply wait for the end of the 'early access' period. Then, they can proceed to download the model for free. Often, even though the model works, it doesn't get any votes. This greatly penalizes the creator who puts effort into sharing the model. As an unsubscribed user, I am penalized by this system. I understand that Civitai has management costs, but it would be nice if even those who collaborate and are not subscribed could receive adequate compensation and an incentive to continue. If possible, please try to donate even a few buzz and always remember to vote for the model. This way, I can continue to share more models with you. Voting for images brings buzz to both the creator and the voter, so it is important to do so. CivitAi's system only gives visibility and more buzz to those who are subscribed. Unfortunately, I can't afford it and rely solely on buzz to be able to train the models. I have had to move many models to Tensor for some reasons.The first reason is that some users on Civitai have downloaded some of my models and images and put them up for sale on Patreon, DeviantArt, and Pixiv. The sad thing is that they have passed them off as their own creations. I reported this, and the result was a ban on my profile. The second reason is how Tensor is structured, providing points for every use of LoRa or Checkpoint even to free users, and these points can be used for training. Additionally, models may have downloads disabled but can still be used for generation and training, giving points to the creator even for free. Some of you have asked me to share them on Civitai, but seeing the ratio between downloads, votes, and buzz, I gave up. For creators, it is sad to see a model downloaded 1,000 times but receiving only 100 votes or less. It seems that the model doesn't work, but that's not the case because the images I publish are created with A1111 on my PC. I don't generate them on Civitai because it costs too much buzz, and the result is not as good as on the PC.
Please forgive the numerous mistakes in my poor English, but I hope you understand. I kindly ask you to always vote for my models and, if you can, also vote for images and videos, as it benefits you too. If possible, please donate a few buzz to my models, because only in this way I can continue to share more models, use training, and create more videos. Otherwise, I will have to give up (unless Civitai changes its mind and returns to the way it used to be). I hope for your understanding. I sincerely thank all those who follow me and have donated buzz, helping me to share my passion.
Now Civitai have implemented the "Auctions". It is good but it is a nonsense because those who have few buzz cannot participate, also in the head there are always the usual ones who have a tide of buzz. So what's the use? Give visibility? Only to spend further buzz who have only those who have been subscribed and who have been at the head of the rankings for months. It is of no use to emerging creators who have few buzz. Instead of inserting something really useful for emerging creators ... they favor, for the umpteenth time, those who do not need it. The new creators is not encouraged. So I will have the models always in the "dormant" state that will cost 40% more and that no one will use anymore, simply because the usual people who are already in the lead have won. Truly a horrible help to new creators... thanks for help us CivitAi.
Some users have asked me why I still create models and LoRAs with SD1.5. This is because it is faster, and with SD I can now generate all kinds of images. SDXL or PONY require more powerful PCs and GPU, and it takes me 10 minutes to generate a single image (i have RTX2060). Moreover, the realism in SDXL or PONY is not much different from what I can achieve with SD1.5, but since SD1.5 is faster, I prefer it. Sure, SD has issues with hands, and it is more difficult to follow a prompt, but you can see the results in my gallery images.
Others have asked how I create videos, this is possible through the 'image to video' conversion using the KlingAi plugin from CivitAi. I use CFG at 6 and simply write what I want to happen in the video in the prompt. Example, "the girl walk", "wind move her hair", etc.
I add, for those who say my model does not work or works poorly, I must remind that my model or a lora that requires trigger words, these words must be used. If they are not used, the loras or models will not function. Moreover, a prompt with many loras can cause malfunctions to the main loras because some are not compatible with each other. I recommend using my loras alone, as they are optimized to work independently. My optimizers are included in the loras and Enhanced models, so they should not be used. To get images like mine, just use the recommended positive and negative prompts and all the trigger words indicated in the model/lora.
I thank in advance anyone who will donate some buzz to allow me to continue and train new Loras and models. Remember to vote always for images and models.
VERY IMPORTANT: As explained above, to improve the visibility of a model, it depends on you. Situations like these indicated in the photo, where a model is downloaded more than 2000 times, but it is voted only by 200 people, it penalizes a lot those who created the model. I ask you to always vote for the model, otherwise it is useless to continue sharing the creations on this site. If you have problems using a model, follow the recommended settings and you will see that it will work. Alternatively, write me on the model page. Thank you.
NEWS 15 march : After the last update of CivitAi, many lora or models work poorly, waiting for more positive developments, I think I will publish the other NEW models and lora on the TA website. At least there they work correctly because during the generation you can choose A1111 or ComfUi while on Civitai this is not possible. In addition, the generation has become very bad. This is due to Civitai's incorrect upgrade, not my models working properly with A1111 prompts. If civitai will fix this mess, I will insert all the new lora and models.
NEWS 18 march: Fortunately, Civitai rethought it, I hope there will be no more similar upgrades.
NEWS 28 march : Since civitai has included the use of checkpoints again, according to a ranking based on how many buzz are used, my checkpoints can never be used. For this reason, it is useless to publish my checkpoints since they will never be used. You can use them for free generation and training on Tensor Art.
APRIL : Given the few comments (without comments I can't know if the model likes or has problems) in the models (many put the emoticon that cries I don't know why) and the low participation and positive votes, unfortunately I will find myself forced to no longer support CivitAi. All new checkpoints will be posted only on TensorArt. Those who want to can try them out and also use them for training in TA site. For the moment I will only publish the LORAs, as long as I have buzz available. I'm sorry about this decision, but I don't have the hardware and resources to compete with the recent (illustrious) models and creator with subscription. Also civitai does not allow me to be able to highlight the models without a subscription. Greetings, follow me on Tensor Art.
Bye, have a good day and creations.