clamPDXL (Find&ForgetYou PonyXL)

Welcome to clamPDXL, an attempt to port the art style of my Find&ForgetYou and Find&ForgetYou – Anime series of checkpoints to PonyXL. Trained on a pruned dataset of imaged generated on these checkpoints and uploaded to CivitAI (so really, this is your fault as much as it is mine), and merged with AutismMix_confetti.

While my goal was to try and fill the void of semi-realistic PonyXL chackpoints on CivitAI, what I ended up with was a versatile model capable of both the classic Find&ForgetYou semi-realism as well as anime, with the help of PonyDiffusion specific prompting. Check the gallery above for an xyz grid, read the description of PonyDiffusionV6 for a more in-depth explanation of the tags, or just keep reading for a basic overview of suggested prompts on this model.

Classic Find&ForgetYou Semi-Real Prompt

positive:

score_9, score_8_up, score_7_up, {prompt}

negative:

Semi-Real Variations
positive:

score_9, score_8_up, score_7_up, source_anime, {prompt}

negative:

or

positive:

score_9, score_8_up, score_7_up, source_anime, {prompt}

negative:

source_cartoon, source_pony, source_furry

Fuck You, Give Me Anime

positive:

score_9, score_8_up, score_7_up, source_anime, {prompt}

negative:

3d

or

positive:

score_9, score_8_up, score_7_up, source_anime, {prompt}

negative:

source_cartoon, source_pony, source_furry, 3d

These are just suggestions based on my preliminary testing! I encourage you to mix and match and experiment to find something that suits you!

Sample Image Settings:

• Sampler – Euler a

• Sampling Steps – 24

• Dimensions – 832×1216

• CFG Scale – 7

• Upscaler – 4x-AnimeSharp

• Hires Steps – 12

• ADetailer – face_yolo8n

this checkpoint is terrible and needs a lot more work, but i spent too much time on it to not take up 6gb of your ssd