This model is able to generate anthro characters as well as human characters with minimal prompting. It also is able to work well with various LoRAs. Being a merge of merges created using the original Pony Diffusion V6 model, it is advised to use the "score_9, score_8_up, score_7_up," prefix at the beginning of all prompts, though some prompts omitting the prefix still yield coherent images. See the image showcase for examples.
This model was created by merging two merged checkpoints:
The first merge was a 0.4 weighted sum between Perfect NAI V11 and Prefect Pony XL V3.
The second merge was a 0.4 weighted sum between Fantasm Pony XL V10 and waiANINSFWPONYXL_v70.
The resulting checkpoints were merged together in a final 0.4 weighted sum.
