If you like what I do, leave me a ❤️ to receive notifications of future updates and follow me here on Civitai. Remember that it is a hobby for me and it costs a lot of time and everything is shared entirely for free, I appreciate that you make a review and share your creations on this wall.
Armor from HaDeS XL v4.0 ReTRaiNeD
Completely retrained from scratch with a new dataset.
Responds well to simple and complex prompts.
The idea of this retraining is to add more details and implement an armor selector in future updates based on this model as a base.
Armor from HaDeS XL v3.4 / v3.4 Lite
After a few months developing a checkpoint with good quality to improve this LoRa (see Realism from HaDeS XL), I can finally release this update that greatly improves previous versions. I hope you like it and that you share your images with the community.
As I understand that many of you work with many LoRas, I release a much lighter version (v3.4 Lite) that saves much more disk space.
Please read the following carefully:
This LoRa has been trained to integrate perfectly into Realism from HaDeS XL v3.5 EVO, which means it also works perfectly in Realism from HaDeS XL Lightning. It may work well in other checkpoints or not, it has not been fully tested outside the ecosystem of the checkpoints of my authorship.
Armor from HaDeS XL v2.1
Added Spiderman armor with trigger word: rmspdvrs
P.S.: add "spider logo" to the prompt if needed.
Armor from HaDeS XL v2.0
introducing new armors and variations. Each one is activated with its own trigger word:
nhdsrmr: It is the main armor of the set, a direct variation of the original Armor from HaDeS.
chhdsrmr: A chromatic armor, perfect to mix with the others as it adds reflections and softens the lines of the armor.
hlhdsrmr: Add crystals, more pointed shapes, cooler colors
crstlhdsrmr: It is the new armor evolution of Armor from HaDeS and will be the base of future updates, it adds more details.
Each trigger word can be used separately or combined with one, two, three, or all together.
I hope you enjoy this LoRa, have a lot of fun creating amazing armors with it, and give it a lot of love and share all your creations on this wall.
Cya in the next update!
Armor from HaDeS XL v1.0
HaDeS Armor comes to Stable Diffusion XL with a simple but very attractive first version. I hope you like it and have fun with this model. Don't forget to share your armor pictures with the entire community and give it a lot of love and feedback to bring you an even better V2!!
if you like my work and want to support me, or are interested in information about my upcoming projects or updates to my models, visit: https://ko-fi.com/devildonia.
Description
v1.0
FAQ
Comments (2)
I haven't fully explored the capabilities of LORA yet, but I have a suggestion. The abundance and quality of details are immense, almost overwhelming. Looking at such a large number of details, the eye needs some space to rest, meaning larger elements with fewer details. This makes the composition more enjoyable to perceive, allowing the eye to pause and relax. It seems to me that adding more larger, simple elements would add a lightness to the overall composition. My eye is simply craving more flat surfaces, something like that orange armor where you can see solid elements on the chest plate. I understand that finding a balance between such complex elements while training LORA can be challenging.
I noticed that there is only one tag "hdsrmr" in the model. Is there an option to train a model that can generate two armors with different levels of detail? What if you used two different sets, like "High Detailed" and "Low Detailed," for training, allowing control in the prompt to specify which type to generate? I sometimes try to train my own models and am considering such solutions.
Nevertheless, great job! Keep it up. I look forward to eagerly awaiting the next models. Cheers!
First of all, thank you for your very constructive and detailed comment. It is a pleasure to meet users who thoroughly test your LoRa and make you see the good and the bad of it.
Armor from HaDes XL is still in its first version. I wanted to see how far Stable Diffusion XL was able to go in terms of details. And I have been pleasantly surprised by the high quality you can achieve with a well-trained LoRa. Now the next step will be to add more different armors as in its namesake from Stable Diffusion 1.5. The idea of adding different qualities to the armor can be easily done by lowering the strength of the model, so it shouldn't be difficult to implement (at least on paper, then when it comes to training anything can happen xD).
Thank you very much again for your time, that helps me a lot to improve my LoRas, and see you soon in v2.0