<h1 id="pastel-pinup-palooza-r7">Pastel Pinup Palooza r7</h1><ul><li>Trigger word: kkpinup</li><li>Rank / Alpha: 32</li><li>Training steps: 3,000</li><li>License: Non-commercial</li></ul><h2 id="overview">Overview</h2>Pastel Pinup Palooza r7 is a LoRA trained on a curated corpus of 2,000 pinup-style images spanning the 1930s through the 1990s, covering a wide range of eras, genres, poses, wardrobes, and registers (SFW, NSFW, and X-rated). The training set was produced by my Pinup_Girls pipeline — a structured vocabulary-driven render system that pairs era-authentic styling with diverse figure forms, lighting recipes, and compositional poses.The model captures the classic pinup aesthetic: pastel graphic color palettes, period-authentic fashion and styling, flattering figure-forward compositions, and the playful, glamorous energy of mid-century illustration and photography. Whether you're generating cheesecake glamour, wartime magazine covers, or boudoir scenes, Pastel Pinup Palooza aims to deliver consistent, high-quality pinup imagery with strong era fidelity.<h2 id="intent-and-design-goals">Intent &amp; Design Goals</h2><ul><li>Era fidelity — Each era (1930s–1990s) carries a distinct palette, anchor phrasing, and material-culture cues (victory rolls, tailored suits, print fabrics, etc.) that the model has learned as strong conditioning signals.</li><li>Pose diversity — 70+ distinct pose descriptions covering seated, standing, reclining, and action poses, ensuring compositional variety.</li><li>Register flexibility — Trained across SFW (cheesecake, glamour), NSFW (lingerie, swimwear), and X-rated (nude/boudoir) registers so the user can control content level through prompt wording.</li><li>Hair as a promptable attribute — Hair descriptions are included as a required caption clause, so specific hairstyles can be requested or varied.</li><li>Palette conditioning — Era-appropriate color palettes are embedded in captions; specifying a palette in your prompt will steer the color treatment.</li></ul><h2 id="how-to-prompt">How to Prompt</h2><h3 id="basic-usage">Basic usage</h3>Include the trigger word kkpinup in your prompt. It activates the pinup style. Add the phrase pinup illustration to your prompt to activate the pastel pinup illustration style.<pre><code>kkpinup, 1950s cheesecake, pinup illustration of a woman in polka-dot swimsuit, seated on beach chair, sunny lighting</code></pre><h3 id="era-control">Era control</h3>Specify the decade to steer the styling, palette, and material culture:<ul><li>1930s — Art deco, muted tones, bias-cut silhouettes</li><li>1940s — Wartime magazine illustration, victory rolls, tailored suits</li><li>1950s — Cheesecake glamour, bold color, polka dots, pin curls</li><li>1960s — Mod fashion, bold geometric patterns, bouffant hair</li><li>1970s — Disco era, flared silhouettes, earth tones, feathered hair</li><li>1980s — Neon accents, bold shoulders, big hair, high-contrast color</li><li>1990s — Grunge minimalism, slip dresses, muted palettes, effortless styling</li></ul><h3 id="register-content-level">Register / content level</h3>Control the content register through descriptive wording:<ul><li>SFW: cheesecake, glamour, swimsuit, sundress, lingerie (covered)</li><li>NSFW: lingerie, boudoir, unbuttoned, sheer, revealing</li><li>X-Rated: nude, topless, bottomless, full nudity, boudoir scene</li></ul>The model was trained on all registers. The output register is determined by your prompt — be explicit about what you want.<h3 id="pose-and-composition">Pose &amp; composition</h3>Use pose descriptors to control composition. Examples from the training vocabulary:<ul><li>seated on a low stool, at her vanity, one hand lifting the hair from her neck</li><li>standing, hands on hips, weight on one leg, looking over shoulder</li><li>reclining on a chaise, legs crossed, arm behind head</li><li>leaning against a car fender, arms crossed, confident smile</li><li>perched on a bar stool, legs bent on rungs, eyes to camera</li></ul><h3 id="hair-control">Hair control</h3>Hair is a promptable attribute. Specify a style, for example:<ul><li>victory rolls</li><li>pin curls</li><li>bouffant</li><li>pageboy</li><li>long waves</li><li>updo</li><li>ponytail</li><li>finger waves</li></ul><h3 id="palette-control">Palette control</h3>Steer the color treatment by naming a palette:<ul><li>print inks — cherry red, warm gold, cream, slate navy (1940s)</li><li>saturated wartime print inks (early 1940s)</li><li>pastel palette, muted tones, bold primary colors</li></ul><h3 id="putting-it-together-example-prompts">Putting it together — example prompts</h3>SFW cheesecake:<pre><code>kkpinup, 1950s, cheesecake, pinup illustration featuring a woman in a polka-dot sundress, standing with hands on hips, pin curls, bright sunny lighting, bold color palette, eyes to camera</code></pre>NSFW boudoir:<pre><code>kkpinup, 1940s, boudoir, pinup illustration featuring a woman in lace lingerie, seated on a low stool at her vanity, one hand lifting hair from neck, victory rolls, muted wartime print inks, soft lighting</code></pre>Full example with pose:<pre><code>kkpinup, 1960s, mod fashion, pinup illustration featuring a woman in geometric print mini dress, reclining on chaise, legs crossed, arm behind head, bouffant hair, bold primary colors, studio lighting</code></pre><h3 id="using-the-embedded-workflow-in-sample-images">Using the embedded workflow in sample images</h3>The sample images included with this model have a ComfyUI workflow embedded in their PNG metadata. You can use these as a ready-made starting point:<ol><li>Drag and drop any sample image into ComfyUI — the full workflow (nodes, prompts, LoRA loading, and settings) will load automatically. You will need to update the locations of the models and the LoRA in the UI.</li><li>Modify the prompt in the text nodes to suit your needs — swap era, pose, hair, palette, and register descriptors as described above.</li><li>Adjust the LoRA weight in the LoRA loader node (start at 0.8–1.0) to control style strength.</li><li>Swap the base model if you're using a different checkpoint — the workflow structure remains the same. The sample images were rendered with the bf16 versions. If you have low VRAM, update to the int8 versions.</li><li>Add or remove nodes to integrate the LoRA into your existing pipeline (e.g., adding upscalers, ControlNet, or additional LoRAs).</li></ol>Tip: The embedded workflow uses the same node layout and parameter set that produced the training corpus, so it's already tuned for the kkpinup trigger word and the structured caption format. Use it as a baseline, then iterate.<h2 id="tips-and-best-practices">Tips &amp; Best Practices</h2><ul><li>Always include kkpinup — without the trigger word, the style won't activate.</li><li>Add the phrase pinup illustration to get the trained pastel pinup illustration — without the trigger phrase, the style will vary and lean toward photo-realism.</li><li>Be specific about era — the model responds strongly to decade keywords; vague prompts may blend eras.</li><li>Layer descriptors — combine era + genre + pose + hair + palette for the most controlled results.</li><li>Register follows wording — if you want SFW output, avoid nudity-related terms; the model respects your prompt for the most part.</li><li>Negative prompts — if you get unwanted modern elements, try: modern, contemporary, anachronistic, digital, 3d render, photograph</li><li>Strength — start at LoRA weight 0.8–1.0; reduce to 0.6–0.7 for subtler style influence.</li></ul><h2 id="limitations">Limitations</h2><ul><li>Era blending — Without explicit era keywords, the model may produce a generic pinup look that blends decades.</li><li>Anatomy — As with all diffusion models, extreme poses or complex hand positions may produce artifacts.</li><li>Register control — The model is trained on all registers; if you want SFW-only output, avoid nudity-adjacent terms in your prompt.</li><li>Color saturation — 1940s palettes were trained with both saturated and muted variants; results may vary.</li></ul><h2 id="credits-and-license">Credits &amp; License</h2><ul><li>Training corpus: The 650 image training set for this model was selected from a set of 2,000 curated images generated by my Pinup_Girls pipeline.</li><li>Model name: Pastel Pinup Palooza r7</li><li>Artistic License: This LoRA was trained using a corpus of images generated randomly by multiple diffusion models. Any resemblance to a specific artist style is purely coincidental.</li><li>Non-commercial license: This model is released for personal, educational, and non-commercial use only. Commercial use requires explicit permission from the model author. Don't make me wake my lawyer. Without her beauty sleep, she is cranky, and you will pay the price.</li></ul>

Pastel Pinup Palooza (Moss SFT | Qwen Image 2.1 | Z-Image Turbo | Chroma.1 HD | Krea 2 Turbo | Flux.2 Klein 9b )

MOSSImageSFT_PastelPinup_kkpinup_r7.safetensors

MOSS Image SFT r7

Files with SHA256: 65d1efd9e4492bb5f326bd70b3ca6bb734deb9e66c0172cdb85eba6131d04f53