This is a first version of a model trained in SD2.1, with a dataset of 3+k images spanning from high-res images of paintings of famous artists down to high-res pictures of single hands, faces, landscapes, and so on.
Trained on large resolution images for over 35 hours, it is able to produce 1200x900 images right out of the box, covering almost all art styles and topics.
Some things it has not seen (such as naked bodies), so forget it if you want to make nudes
This is an art station, that is capable of producing perhaps not as anatomically correct as one would wish for true paintings and art :)
Be creative, see the images provided for some sample prompts, and be not confused as I managed again to rename the model after generating the first sample images, so some of the prompts state to be made model that is not named tgl_artstation_ep49
Yet, it is the same model, as you will be able to confirm when running it
Big thanks to "followfox" for the public posts on EveryDream, which pushed me away from DreamBooth towards EveryDream, a decision I will not regret.
描述:
训练词语:
名称: tglArtstationV1_v10.ckpt
大小 (KB): 2519886
类型: Model
Pickle 扫描结果: Success
Pickle 扫描信息: No Pickle imports
病毒扫描结果: Success
