Most Stable Diffusion checkpoints have a rough idea of what doomguy looks like. But they mix all versions and variations together. This LoRA was trained on 30 fanarts of the classic doomguy from the 1993 Doom video game over AnyLoRA. It should works on most SD 1.5 checkpoints.
In addition to screenshot from games by idSoftware and Bethesda, training was done using artworks by the following artists: Dakota Lee, TheDOomMan, @springhakurei on Twitter (deleted account), TsitrA, Dmitriy Druzhkov, Rusklav, Yare-Yare-Dong, Heroes Workshop, SignalFire_Plae, KelvinHiu, Dakota DeGarmo, hundsdorfex, Ephssis, Sergey ( TrueMakaR ) Makarich, bonnieta123, night_honoki, Nathanael Julian, Olga Malofeeva, Roddy E, Dekanai, Sherif Saad, and GoblinZac. See here for details.