Modelpublic

ij/anima-humanize-composition-r64-v1

sourceHugging Faceotherupdated 3mo agoView on Hugging Face

0likes54downloads

Model Card

Anima Humanize Composition LoRA r64 v1

An experimental LoRA for circlestone-labs/Anima.

It is intended to make Anima outputs easier to use as character-focused cover illustrations: clearer focal placement, stronger page balance, and fewer awkward crops.

[image]

Model File

text

anima_humanize_composition_r64_v1.safetensors

Recommended Use

text

LoRA strength: 1.0

Example ComfyUI stack:

text

Base model: Anima
This LoRA: 1.0
Resolution used in examples: 720 x 1024
Sampler used in examples: er_sde, simple scheduler
Steps / CFG used in examples: 28 / 3.0

Prompt Format

Positive prompt:

text

newest, year2024, year 2025, (masterpiece, best quality, score_8:0.85), highres, absurdres, sensitive, safe, novel cover, good composition, detailed background, face focus, (animated, dynamic pose:0.1), @artist_tag_1, @artist_tag_2, @artist_tag_3, @artist_tag_4, @artist_tag_5. cover-derived caption

Negative prompt:

text

(expanded view:1.3), (bad composition, cropped, bad crop:1.3), (ai generated:1.3), comic panel, chibi, (worst quality), low quality, cropped, (score_1), score_2, score_3, unfinished, work-in-progress, blank, letterboxed, (blurry, jpeg artifacts, lowres:1.1) mutated, (extra digit:1.4), artist name, logo, messy, crowded, multiple others, cartoon style, old, year 2012, year 2013, year 2014, year 2015, extra digits, (malformed hands:1.5), sepia, old, oldest, old photo, dull, big black navel, furry, anthro, extra toes,

Example Prompt

Positive:

text

newest, year2024, year 2025, (masterpiece, best quality, score_8:0.85), highres, absurdres, sensitive, safe, novel cover, good composition, detailed background, face focus, (animated, dynamic pose:0.1), @onegingek, @noes, @zhengqi_zhizi_sg, @kingchenxi, @tokunou_shoutarou. A glossy rendering in an ornate style shows a young man seated on a dark couch near a tall window, framed from a slightly elevated three-quarter view with his clasped hands resting above a vertical sword. The scene uses vivid colors, glowing lighting, dynamic effects, and a complex cityscape background with bright sky, clouds, sun flares, and sparkling highlights around the foreground.

Negative:

text

(expanded view:1.3), (bad composition, cropped, bad crop:1.3), (ai generated:1.3), comic panel, chibi, (worst quality), low quality, cropped, (score_1), score_2, score_3, unfinished, work-in-progress, blank, letterboxed, (blurry, jpeg artifacts, lowres:1.1) mutated, (extra digit:1.4), artist name, logo, messy, crowded, multiple others, cartoon style, old, year 2012, year 2013, year 2014, year 2015, extra digits, (malformed hands:1.5), sepia, old, oldest, old photo, dull, big black navel, furry, anthro, extra toes,

Additional example prompts and seeds are included in `examples.json`.

Comparisons

The examples below compare Anima alone against the same prompt and seed with this LoRA at strength 1.0.

[image]

What To Expect

This LoRA tends to:

—Pull the main subject into a clearer cover-ready focal position.
—Favor more deliberate vertical page composition.
—Reduce weak staging or off-balance framing.
—Keep character-focused prompts easier to read at a glance.

It may also:

—Change pose, camera distance, or foreground emphasis more than expected.
—Make the subject more dominant in the image.
—Require seed-by-seed checking for prompts with strict object placement.

License

This LoRA is a derivative work of Anima and follows the licensing constraints of the base model. Please refer to the Anima model page and its CircleStone Labs non-commercial license before use.

This model is intended for research and non-commercial personal experimentation.

Disclaimer

This is an experimental composition-control LoRA. It does not guarantee better results for every prompt, seed, aspect ratio, or sampler.