<游戏AI研究所>"Game AI Research Institute"
About this version
XL version test version: This model adopts the workflow form of Comfy UI. Since it is a test version, four styles are open for everyone to use! In the future, 60 styles of triggers will be gradually opened, and each style corresponds to different trigger words for your reference!
TAG:
gmic_\(3dicon\)
gmic_\(2dguofeng\)
gmic_\(2dXIANTIAO\)
gmic icon_\(xieshi\)
The above images were generated without using any other optimization methods for tagging. We welcome everyone to collaborate with us and make progress together.
模型仅用于游戏项目的AI产出交流和实验
模型作者: 游戏图标研究所
模型讨论组（可以抢先免费体验未发布模型）：
DISCORD：https://discord.gg/njBMYJ7mRF（推荐）
QQ频道：https://pd.qq.com/s/6ekpt1xei?businessType=9（推荐）（可以抢先免费体验未发布模型）
The models of the Game Icon Research Institute have entered into a strategic partnership with C Station, Similar websites are prohibited from copying these models.They are only for personal study and communication purposes. The right to interpret them belongs to civitai & the Icon Research Institute.
Version 2.0 is suitable for creating icons in a 2D style, while Version 3.0 is suitable for creating icons in a 3D style.
The version is not about the newer the better
Description
Version 3.0 is produced based on the premise of LORA. Compared with version 2.0, the 3D texture is stronger, and it is suitable for the production of 3D icons. For the time being, Vincent Map needs to be used with LoRa (without ControlNet) to achieve the effect on the map. We are committed to Vincent Map Realization requires a longer period of research and development
Welcome to support us!
FAQ
Comments (7)
Why is this labelled as being trained when it clearly has a merge of rev-animated? The female character style is identical for both models.
Also I don't want to seem negative but I've tried all your loras now and they produce way better and more accurate results in the latest rev-animated model than they do in this model.
This model is trained by us with rev as the bottom model, so it will produce such an effect,
This model is trained by us using rev as the base model, so it will produce such an effect. We are only using it to complete the works of the icon. The reason for this is the base model of the training.
The training model in the Dream booth requires a base model to work. Our original intention is to make a model that can generate graphs without using LORA. We will use various models as the base model for training. During exploration, this cannot be avoided. If you train the model, you will know! We are also exploring, not a person who knows everything! If the trained model is similar to others and poor, then we will not train, and we will always be unable to move forward,
@ConceptConnoisseur Ok that makes sense. It looks to me like this model is just adding a more "game icon" type art style to the results so wouldn't it be easier to do this the other way around.?
The loras already get good results in rev-animated for generating the specific game icons so instead of training a whole model to produce a certain style, just let rev-animated create the images and then train a lora for the "game icon" style.
@nagopif737meidecncom347 We have also tried your suggestion. Your human body parts are easy to appear in the REV model, so I tried to change his model. If we have human body parts in the process of drawing game icons, it will be a headache. I think we can take various measures. The head of the family, come and make up for our deficiencies, we are not only doing research on realism, including cartoons, two-dimensional, etc. Rev alone is definitely not enough
@nagopif737meidecncom347 We have also tried your suggestion. Your human body parts are easy to appear in the REV model, so I tried to change his model. If we have human body parts in the process of drawing game icons, it will be a headache. I think we can take various measures. The head of the family, come and make up for our deficiencies, we are not only doing research on realism, including cartoons, two-dimensional, etc. Rev alone is definitely not enough