Add 25 LoRA(s): yfg-chatgpt-4o-style-flux-zit-krea2, manga-style-naoki-urasawa, zen-ink-wash-sumi-e, flux2-klein-anything-to-real-characters…

.gitattributes

@@ -469,3 +469,146 @@ minimalist-vector-art/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs
 minimalist-vector-art/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
 minimalist-vector-art/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
 minimalist-vector-art/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+yfg-chatgpt-4o-style-flux-zit-krea2/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+yfg-chatgpt-4o-style-flux-zit-krea2/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+yfg-chatgpt-4o-style-flux-zit-krea2/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+yfg-chatgpt-4o-style-flux-zit-krea2/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+yfg-chatgpt-4o-style-flux-zit-krea2/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+yfg-chatgpt-4o-style-flux-zit-krea2/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+manga-style-naoki-urasawa/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+manga-style-naoki-urasawa/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+manga-style-naoki-urasawa/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+manga-style-naoki-urasawa/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+manga-style-naoki-urasawa/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+manga-style-naoki-urasawa/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+zen-ink-wash-sumi-e/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+zen-ink-wash-sumi-e/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+zen-ink-wash-sumi-e/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+zen-ink-wash-sumi-e/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+zen-ink-wash-sumi-e/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+zen-ink-wash-sumi-e/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+flux2-klein-anything-to-real-characters/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+flux2-klein-anything-to-real-characters/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+flux2-klein-anything-to-real-characters/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+flux2-klein-anything-to-real-characters/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+flux2-klein-anything-to-real-characters/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+flux2-klein-anything-to-real-characters/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+harness-panel-gag/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+harness-panel-gag/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+harness-panel-gag/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+harness-panel-gag/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+harness-panel-gag/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+harness-panel-gag/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+gothic-dark-fantasy-movie/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+gothic-dark-fantasy-movie/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+gothic-dark-fantasy-movie/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+gothic-dark-fantasy-movie/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+gothic-dark-fantasy-movie/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+gothic-dark-fantasy-movie/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+disney-mid-century-animation/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+disney-mid-century-animation/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+disney-mid-century-animation/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+disney-mid-century-animation/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+disney-mid-century-animation/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+disney-mid-century-animation/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+super-hip-waist/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+super-hip-waist/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+super-hip-waist/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+super-hip-waist/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+super-hip-waist/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+super-hip-waist/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+k-slider-imaging-control/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+k-slider-imaging-control/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+k-slider-imaging-control/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+k-slider-imaging-control/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+k-slider-imaging-control/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+k-slider-imaging-control/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+herbst-photo-analog-film/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+herbst-photo-analog-film/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+herbst-photo-analog-film/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+herbst-photo-analog-film/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+herbst-photo-analog-film/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+herbst-photo-analog-film/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+elusarca-s-detail-enhancer-flux-klein-9b/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+elusarca-s-detail-enhancer-flux-klein-9b/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+elusarca-s-detail-enhancer-flux-klein-9b/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+elusarca-s-detail-enhancer-flux-klein-9b/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+elusarca-s-detail-enhancer-flux-klein-9b/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+elusarca-s-detail-enhancer-flux-klein-9b/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+nsgirl-ultrarealistic-lora/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+nsgirl-ultrarealistic-lora/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+nsgirl-ultrarealistic-lora/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+nsgirl-ultrarealistic-lora/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+nsgirl-ultrarealistic-lora/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+nsgirl-ultrarealistic-lora/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ethereal-gothic-elegance/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ethereal-gothic-elegance/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ethereal-gothic-elegance/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ethereal-gothic-elegance/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ethereal-gothic-elegance/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ethereal-gothic-elegance/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+uglifier/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+uglifier/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+uglifier/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+line-drawing-ce/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+line-drawing-ce/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+line-drawing-ce/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+line-drawing-ce/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+line-drawing-ce/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+line-drawing-ce/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+3d-redmond-3d-render-style-for-sd-xl/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+3d-redmond-3d-render-style-for-sd-xl/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+3d-redmond-3d-render-style-for-sd-xl/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+3d-redmond-3d-render-style-for-sd-xl/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+3d-redmond-3d-render-style-for-sd-xl/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+3d-redmond-3d-render-style-for-sd-xl/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+rly-thot-shot-jada-krea-2-ideogram-4-zib-zit-k9b/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+rly-thot-shot-jada-krea-2-ideogram-4-zib-zit-k9b/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ars-niji-style/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ars-niji-style/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ars-niji-style/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ars-niji-style/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ars-niji-style/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+ars-niji-style/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+vintage-drawing-ce/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+vintage-drawing-ce/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+vintage-drawing-ce/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+vintage-drawing-ce/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+vintage-drawing-ce/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+vintage-drawing-ce/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+retro-pop-ce/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+retro-pop-ce/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+retro-pop-ce/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+retro-pop-ce/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+retro-pop-ce/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+retro-pop-ce/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+2000s-crime-thriller-movie/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+2000s-crime-thriller-movie/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+2000s-crime-thriller-movie/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+2000s-crime-thriller-movie/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+2000s-crime-thriller-movie/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+2000s-crime-thriller-movie/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+klein-detail-slider/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+klein-detail-slider/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+klein-detail-slider/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+klein-detail-slider/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+klein-detail-slider/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+klein-detail-slider/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+dark-fantasy-pulp-pinup/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+dark-fantasy-pulp-pinup/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+dark-fantasy-pulp-pinup/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+dark-fantasy-pulp-pinup/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+dark-fantasy-pulp-pinup/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+dark-fantasy-pulp-pinup/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+delicate-whimsy/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+delicate-whimsy/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+delicate-whimsy/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+delicate-whimsy/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+delicate-whimsy/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+delicate-whimsy/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text
+cutecartoon-redmond-cute-cartoon-lora/example_images/01_image.jpeg filter=lfs diff=lfs merge=lfs -text
+cutecartoon-redmond-cute-cartoon-lora/example_images/02_image.jpeg filter=lfs diff=lfs merge=lfs -text
+cutecartoon-redmond-cute-cartoon-lora/example_images/03_image.jpeg filter=lfs diff=lfs merge=lfs -text
+cutecartoon-redmond-cute-cartoon-lora/example_images/04_image.jpeg filter=lfs diff=lfs merge=lfs -text
+cutecartoon-redmond-cute-cartoon-lora/example_images/05_image.jpeg filter=lfs diff=lfs merge=lfs -text
+cutecartoon-redmond-cute-cartoon-lora/example_images/06_image.jpeg filter=lfs diff=lfs merge=lfs -text

2000s-crime-thriller-movie/2000sCrimeThrillerKlein9b.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f3afad0b38a5e9a2acbc1795e5e817838a64cc0bb1b27014fdba81ab1fc8e2b7
+size 82866728

2000s-crime-thriller-movie/info.txt

@@ -0,0 +1,238 @@
+═══════════════════════════════════════════════════════════════════════════════
+  2000S CRIME THRILLER MOVIE
+═══════════════════════════════════════════════════════════════════════════════
+
+───────────────────────────────────────────────────────────────────────────────
+  IDENTITY
+───────────────────────────────────────────────────────────────────────────────
+  Base model        FLUX.2 [klein] 9B
+  Generation mode   not declared by the author
+  Checkpoint        not determined — verify before use
+  Category          Visual style
+  Author            ArsMachina
+  Source            https://civitai.com/models/920569
+  Provenance        CivitAI · model 920569 · version 2743576
+
+───────────────────────────────────────────────────────────────────────────────
+  WEIGHTS
+───────────────────────────────────────────────────────────────────────────────
+  File              2000sCrimeThrillerKlein9b.safetensors
+  Version           Klein9B
+  Size              79.0 MB
+  Format            SafeTensor / unknown precision
+  SHA256            f3afad0b38a5e9a2acbc1795e5e817838a64cc0bb1b27014fdba81ab1fc8e2b7
+  Published         2026-03-27
+
+───────────────────────────────────────────────────────────────────────────────
+  TRIGGER WORDS
+───────────────────────────────────────────────────────────────────────────────
+  › "ArsMovieStill, movie still from a gritty, high-contrast 2000s crime
+    thriller movie"
+
+───────────────────────────────────────────────────────────────────────────────
+  RECOMMENDED SETTINGS
+───────────────────────────────────────────────────────────────────────────────
+  The author did not publish specific settings.
+  No author settings were published. Community baseline: steps Distilled
+  FLUX.2-klein-9B: 4 (HF card example; it is step-distilled to exactly this,
+  and more steps do not refine further). Base FLUX.2-klein-base-9B: 50 in the
+  HF card example and in ComfyUI's benchmark; community practice runs lower,
+  with ~20 for production and 8-12 for fast iteration reported in the CivitAI
+  consistency-LoRA article., cfg Distilled: guidance_scale 1.0 (HF card). It
+  is guidance-distilled and its transformer config sets
+  guidance_embeds=false, so raising CFG degrades rather than improves. Base:
+  guidance_scale 4.0 (HF card); the CivitAI consistency-LoRA article
+  recommends 3.5-4.5 for most edits., sampler Not stated by BFL or by
+  docs.comfy.org, and the official flux2 GitHub repo documents no sampler,
+  scheduler or shift values either. Community consensus is euler with a
+  resolution-dependent shifted-Euler / flow-match schedule (ComfyUI's Flux2
+  workflows use a scheduler that interpolates shift between base_shift and
+  max_shift by output resolution). Start from ComfyUI's downloadable Flux.2
+  klein workflow template rather than a blog. See unverified., resolution
+  1024x1024 in every BFL HF card example and in CivitAI's 9B training recipe
+  ("resolution": 1024). Non-square ratios work but no official list of
+  supported dimensions or a stated alignment multiple was found — see
+  unverified., lora_strength 1.0 is the diffusers/ComfyUI default and what
+  CivitAI applies to training previews (strength 1.0). Published per-LoRA
+  guidance varies and should be read from the LoRA, not assumed: the widely
+  used Klein 9B consistency LoRA's CivitAI article says start at 0.5 within a
+  0.2-1.0 range (0.2-0.4 loose, 0.5-0.7 balanced, 0.8-1.0 maximum structural
+  stability) and warns that too high a strength stops edits from applying;
+  SOLRICKS' Realistic Detail LoRA card says 0.6-0.8 subtle, 0.8-1.0 stronger,
+  start 0.8; the vafipas663 distillation-delta LoRA card warns it "produces
+  static noise at strength > 0.5. And sometimes, you have to go as low as
+  0.1.".
+
+───────────────────────────────────────────────────────────────────────────────
+  ADOPTION & POPULARITY
+───────────────────────────────────────────────────────────────────────────────
+  Downloads               4,547
+  Rating                  100% positive (386/386)
+  Comments                1
+  Published               162 days ago
+  Download rate           3.4/day since release
+
+  Adoption rank           #104 of 150 in this collection  (Modest)
+  Momentum rank           #98 of 150 by download rate  (Long tail)
+
+  Adoption tier is the percentile of total downloads within this collection;
+  momentum is the percentile of downloads-per-day since release. Momentum is
+  ranked rather than measured against a fixed rate, because this base model
+  is itself new and every LoRA looks fast on an absolute scale. Both numbers
+  are CivitAI's, read on the scrape date at the foot of this file.
+
+───────────────────────────────────────────────────────────────────────────────
+  CHAINING / STACKING
+───────────────────────────────────────────────────────────────────────────────
+  Stacks with others      not stated
+
+  The author gave no chaining guidance. The rules below are the
+  general FLUX.2 [klein] 9B ones, not anything specific to this LoRA.
+
+  General FLUX.2 [klein] 9B rules that apply here (see REFERENCE.txt):
+
+  • Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model
+    output into model input. There is no CLIP side to route — klein LoRAs
+    are model-only because the Qwen3 text encoder is not trained (CivitAI's
+    recipe fixes trainTextEncoder: false).
+  • Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B
+    LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden,
+    5+20 vs 8+24 blocks), not a quality question.
+  • Mixing a "Flux.2 Klein 9B" LoRA and a "Flux.2 Klein 9B-base" LoRA in
+    one stack is fine — those strings describe the training checkpoint, not
+    incompatible tensor layouts. Set the sampler for the checkpoint
+    actually loaded.
+  • Run base-trained LoRAs on the distilled checkpoint when you want speed:
+    HF's klein LoRA guide reports the adapter loads there and typically
+    gives better results than on base, at 4 steps.
+  • A base-to-turbo / distillation-extraction LoRA is a whole-model delta,
+    not a content LoRA. Put it on the base checkpoint only, and treat its
+    strength as a steps/CFG dial rather than a style knob. Do not add it on
+    top of the already-distilled checkpoint.
+  • Read each LoRA's own published strength range instead of defaulting
+    everything to 1.0 — klein LoRAs in the wild ship very different ones
+    (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA,
+    under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).
+  • With a consistency or edit LoRA in the stack, excessive strength
+    suppresses the prompt: the consistency LoRA's article says that if
+    edits stop applying, lower it before changing anything else.
+  • Prefer runtime LoRA patching over baking a LoRA into the checkpoint:
+    ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein
+    models (4B or 9B) since a March 2026 update — issue #13637, still open.
+  • The author never states which weight family this targets. Two different
+    questions, two different answers. Within 9B (distilled vs base — the
+    two CivitAI strings): not a real incompatibility. Both load through the
+    same Flux2KleinPipeline and have the same repo layout; for the 4B pair,
+    where both transformer configs are public, the distil
+
+───────────────────────────────────────────────────────────────────────────────
+  PROS
+───────────────────────────────────────────────────────────────────────────────
+  + Well established — 4,547 downloads. [source: CivitAI stats]
+  + Positively received — 386 thumbs-up with no down-votes. [source:
+    CivitAI stats]
+  + Explicit trigger word(s) — ArsMovieStill, movie still from a gritty,
+    high-contrast 2000s crime thriller movie — so the effect can be turned
+    on and off from the prompt. [source: model metadata]
+  + Compact at 79 MB — cheap to stack with other LoRAs. [source: file size]
+
+───────────────────────────────────────────────────────────────────────────────
+  CONS & CAVEATS
+───────────────────────────────────────────────────────────────────────────────
+  - Author never states which generation mode it was trained for; test
+    against your own workflow before relying on it. [source: description]
+  - Thin documentation — expect trial and error. [source: description
+    length]
+
+───────────────────────────────────────────────────────────────────────────────
+  COMPATIBILITY WARNING
+───────────────────────────────────────────────────────────────────────────────
+  Two different questions, two different answers. Within 9B (distilled vs
+  base — the two CivitAI strings): not a real incompatibility. Both load
+  through the same Flux2KleinPipeline and have the same repo layout; for the
+  4B pair, where both transformer configs are public, the distilled and base
+  configs are byte-for-byte identical in every shape field, and the 9B pair
+  is built the same way. HF's own klein LoRA guide states the workflow
+  directly: "Distilled models are step-compressed for fast inference; you
+  train against the base checkpoint and the adapter still loads on the
+  distilled model afterward", and "Applying the LoRA on the distilled model
+  typically gives better results than the base model, and it's faster" (that
+  post is written against 4B, but the family structure is the same). So the
+  two CivitAI strings record which checkpoint the author trained or previewed
+  against, not a hard compatibility wall. The real risk is a silent settings
+  mismatch: load a base-trained LoRA on the distilled checkpoint while
+  keeping the author's 20-50 step / CFG ~4 preview settings and the output
+  degrades badly (distilled wants 4 steps / guidance 1.0), or the reverse.
+  Nothing errors — it just looks wrong. Across sizes (9B vs 4B): a genuine
+  architectural mismatch, from the published transformer configs — different
+  hidden width (4096 vs 3072), different block counts (8+24 vs 5+20) and a
+  different text-conditioning dim (12288 vs 7680, following Qwen3-8B vs
+  Qwen3-4B). A 4B LoRA cannot patch a 9B model: expect missing keys and shape
+  errors, not degraded output. The exact ComfyUI error text is unverified. KV
+  variant: same 9B size and same 4-step distillation, but a distinct
+  checkpoint with its own pipeline; whether standard klein 9B LoRAs patch it
+  cleanly is not documented — see unverified.
+
+───────────────────────────────────────────────────────────────────────────────
+  AUTHOR'S DESCRIPTION
+───────────────────────────────────────────────────────────────────────────────
+  A LoRA inspired by the 2000s crime thrillers.
+
+  The trigger words are ArsMovieStill, movie still from a gritty, high-
+  contrast 2000s crime thriller movie
+
+───────────────────────────────────────────────────────────────────────────────
+  EXAMPLE MEDIA
+───────────────────────────────────────────────────────────────────────────────
+  6 preview(s) in ./example_images/ — community results for
+  this LoRA. Videos keep a matching _poster.jpg still frame.
+
+  No prompts are recorded below: CivitAI's public API flags these posts as
+  having generation metadata but does not return it, so the prompts cannot be
+  scraped. Open the model page to read them.
+
+  01_image.jpeg  (image, 1024x1024)
+  02_image.jpeg  (image, 1024x1024)
+  03_image.jpeg  (image, 1024x1024)
+  04_image.jpeg  (image, 1920x1024)
+  05_image.jpeg  (image, 1920x1024)
+  06_image.jpeg  (image, 1920x1024)
+
+───────────────────────────────────────────────────────────────────────────────
+  HOW TO USE (COMFYUI)
+───────────────────────────────────────────────────────────────────────────────
+  1. Diffusion model -> ComfyUI/models/diffusion_models/. Distilled:
+  flux-2-klein-9b-fp8.safetensors. Base: flux-2-klein-
+  base-9b-fp8.safetensors. (BF16 single-file weights
+  flux-2-klein-9b.safetensors / flux-2-klein-base-9b.safetensors from the
+  gated HF repos also work if you have ~29GB; community GGUF builds exist
+  from unsloth and leejet for lower-VRAM setups.)
+  2. Text encoder ->
+  ComfyUI/models/text_encoders/qwen_3_8b_fp8mixed.safetensors (from Comfy-
+  Org/vae-text-encorder-for-flux-klein-9b). This is the 9B family's encoder;
+  the 4B family uses qwen_3_4b.safetensors — do not mix them.
+  3. VAE -> ComfyUI/models/vae/flux2-vae.safetensors (shared across klein
+  sizes).
+  4. LoRA -> ComfyUI/models/loras/ as a plain .safetensors.
+  5. Load with Load Diffusion Model (UNETLoader) for the transformer,
+  CLIPLoader for the Qwen3 text encoder, VAELoader for the VAE.
+  docs.comfy.org ships prebuilt Flux.2 klein graphs under Workflow Templates
+  for both the distilled and base branches — start from those rather than
+  wiring by hand, since neither the docs nor BFL's repo state
+  sampler/scheduler/shift in prose; those values live in the template.
+  6. Insert LoraLoaderModelOnly between the diffusion-model loader and the
+  sampler. Model-only is correct for klein: the Qwen3 text encoder is not
+  trained by the standard klein trainers, so there is no text-encoder half to
+  patch.
+  7. Set steps and guidance to match the checkpoint you actually loaded, not
+  the one the LoRA was trained on: 4 steps / guidance 1.0 on distilled,
+  ~20-50 steps / guidance ~4.0 on base.
+  8. Note the 9B repos are gated on HuggingFace — you must accept the FLUX
+  Non-Commercial License on the repo page before any download, including
+  scripted ones.
+
+───────────────────────────────────────────────────────────────────────────────
+  Scraped by lorakit on 2026-09-05 01:52 UTC.
+  Stats and description are the author's; pros/cons are derived from the
+  evidence tagged beside each line.
+───────────────────────────────────────────────────────────────────────────────

2000s-crime-thriller-movie/metadata.json

@@ -0,0 +1,134 @@
+{
+  "name": "2000s Crime Thriller Movie",
+  "slug": "2000s-crime-thriller-movie",
+  "base_model": "FLUX.2 [klein] 9B",
+  "media_type": "image",
+  "model_id": 920569,
+  "version_id": 2743576,
+  "version_name": "Klein9B",
+  "source": "civitai",
+  "url": "https://civitai.com/models/920569",
+  "repo_id": null,
+  "file_path": null,
+  "author": "ArsMachina",
+  "category": {
+    "key": "style",
+    "label": "Visual style"
+  },
+  "modalities": [],
+  "primary_modality": "unspecified",
+  "checkpoint_family": "unknown",
+  "trigger_words": [
+    "ArsMovieStill, movie still from a gritty, high-contrast 2000s crime thriller movie"
+  ],
+  "recommended_settings": {},
+  "chaining": {
+    "author": {
+      "stacks": "not stated",
+      "notes": "",
+      "avoid": [],
+      "requires": [],
+      "order": "",
+      "strength": "",
+      "stated": false
+    },
+    "general_rules": [
+      "Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model output into model input. There is no CLIP side to route — klein LoRAs are model-only because the Qwen3 text encoder is not trained (CivitAI's recipe fixes trainTextEncoder: false).",
+      "Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden, 5+20 vs 8+24 blocks), not a quality question.",
+      "Mixing a \"Flux.2 Klein 9B\" LoRA and a \"Flux.2 Klein 9B-base\" LoRA in one stack is fine — those strings describe the training checkpoint, not incompatible tensor layouts. Set the sampler for the checkpoint actually loaded.",
+      "Run base-trained LoRAs on the distilled checkpoint when you want speed: HF's klein LoRA guide reports the adapter loads there and typically gives better results than on base, at 4 steps.",
+      "A base-to-turbo / distillation-extraction LoRA is a whole-model delta, not a content LoRA. Put it on the base checkpoint only, and treat its strength as a steps/CFG dial rather than a style knob. Do not add it on top of the already-distilled checkpoint.",
+      "Read each LoRA's own published strength range instead of defaulting everything to 1.0 — klein LoRAs in the wild ship very different ones (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA, under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).",
+      "With a consistency or edit LoRA in the stack, excessive strength suppresses the prompt: the consistency LoRA's article says that if edits stop applying, lower it before changing anything else.",
+      "Prefer runtime LoRA patching over baking a LoRA into the checkpoint: ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein models (4B or 9B) since a March 2026 update — issue #13637, still open.",
+      "The author never states which weight family this targets. Two different questions, two different answers. Within 9B (distilled vs base — the two CivitAI strings): not a real incompatibility. Both load through the same Flux2KleinPipeline and have the same repo layout; for the 4B pair, where both transformer configs are public, the distil"
+    ]
+  },
+  "pros": [
+    "Well established — 4,547 downloads. [source: CivitAI stats]",
+    "Positively received — 386 thumbs-up with no down-votes. [source: CivitAI stats]",
+    "Explicit trigger word(s) — ArsMovieStill, movie still from a gritty, high-contrast 2000s crime thriller movie — so the effect can be turned on and off from the prompt. [source: model metadata]",
+    "Compact at 79 MB — cheap to stack with other LoRAs. [source: file size]"
+  ],
+  "cons": [
+    "Author never states which generation mode it was trained for; test against your own workflow before relying on it. [source: description]",
+    "Thin documentation — expect trial and error. [source: description length]"
+  ],
+  "file": {
+    "name": "2000sCrimeThrillerKlein9b.safetensors",
+    "size_bytes": 82866728,
+    "size_mb": 79.03,
+    "sha256": "f3afad0b38a5e9a2acbc1795e5e817838a64cc0bb1b27014fdba81ab1fc8e2b7"
+  },
+  "stats": {
+    "downloads": 4547,
+    "thumbs_up": 386,
+    "thumbs_down": 0
+  },
+  "popularity": {
+    "downloads": 4547,
+    "thumbs_up": 386,
+    "thumbs_down": 0,
+    "comments": 1,
+    "approval": "100% positive (386/386)",
+    "age_days": 161.7,
+    "downloads_per_day": 3.36,
+    "rank_by_downloads": 104,
+    "rank_by_momentum": 98,
+    "of_total": 150,
+    "adoption_tier": "Modest",
+    "momentum": "Long tail"
+  },
+  "selection_score": 9179.0,
+  "selection_reasons": [
+    "4,547 downloads",
+    "386 thumbs-up",
+    "top-ranked for its category"
+  ],
+  "example_media": [
+    {
+      "file": "01_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "02_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "03_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "04_image.jpeg",
+      "type": "image",
+      "resolution": "1920x1024",
+      "prompt": ""
+    },
+    {
+      "file": "05_image.jpeg",
+      "type": "image",
+      "resolution": "1920x1024",
+      "prompt": ""
+    },
+    {
+      "file": "06_image.jpeg",
+      "type": "image",
+      "resolution": "1920x1024",
+      "prompt": ""
+    }
+  ],
+  "tags": [
+    "style",
+    "still",
+    "midjourney v6",
+    "movie"
+  ],
+  "scraped_at": "2026-09-05T01:52:09+00:00",
+  "path": "flux2-klein-9b/2000s-crime-thriller-movie"
+}

3d-redmond-3d-render-style-for-sd-xl/[FLUX.2.Klein]3DRenderStyle_Redmond.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d364fc6eecd48d5c5eebb3f427b42596457903598005ae70ee73056de8a4fa83
+size 165704504

3d-redmond-3d-render-style-for-sd-xl/info.txt

@@ -0,0 +1,260 @@
+═══════════════════════════════════════════════════════════════════════════════
+  3D.REDMOND - 3D RENDER STYLE FOR SD XL
+═══════════════════════════════════════════════════════════════════════════════
+
+───────────────────────────────────────────────────────────────────────────────
+  IDENTITY
+───────────────────────────────────────────────────────────────────────────────
+  Base model        FLUX.2 [klein] 9B
+  Generation mode   not declared by the author
+  Checkpoint        not determined — verify before use
+  Category          Visual style
+  Author            artificialguybr
+  Source            https://civitai.com/models/194138
+  Provenance        CivitAI · model 194138 · version 2705473
+
+───────────────────────────────────────────────────────────────────────────────
+  WEIGHTS
+───────────────────────────────────────────────────────────────────────────────
+  File              [FLUX.2.Klein]3DRenderStyle_Redmond.safetensors
+  Version           v1.0 - Flux klein 9b
+  Size              158.0 MB
+  Format            SafeTensor / unknown precision
+  SHA256            d364fc6eecd48d5c5eebb3f427b42596457903598005ae70ee73056de8a4fa83
+  Published         2026-02-20
+
+───────────────────────────────────────────────────────────────────────────────
+  TRIGGER WORDS
+───────────────────────────────────────────────────────────────────────────────
+  › "3DRenderAF"
+  › "3D Render Style"
+
+───────────────────────────────────────────────────────────────────────────────
+  RECOMMENDED SETTINGS
+───────────────────────────────────────────────────────────────────────────────
+  The author did not publish specific settings.
+  No author settings were published. Community baseline: steps Distilled
+  FLUX.2-klein-9B: 4 (HF card example; it is step-distilled to exactly this,
+  and more steps do not refine further). Base FLUX.2-klein-base-9B: 50 in the
+  HF card example and in ComfyUI's benchmark; community practice runs lower,
+  with ~20 for production and 8-12 for fast iteration reported in the CivitAI
+  consistency-LoRA article., cfg Distilled: guidance_scale 1.0 (HF card). It
+  is guidance-distilled and its transformer config sets
+  guidance_embeds=false, so raising CFG degrades rather than improves. Base:
+  guidance_scale 4.0 (HF card); the CivitAI consistency-LoRA article
+  recommends 3.5-4.5 for most edits., sampler Not stated by BFL or by
+  docs.comfy.org, and the official flux2 GitHub repo documents no sampler,
+  scheduler or shift values either. Community consensus is euler with a
+  resolution-dependent shifted-Euler / flow-match schedule (ComfyUI's Flux2
+  workflows use a scheduler that interpolates shift between base_shift and
+  max_shift by output resolution). Start from ComfyUI's downloadable Flux.2
+  klein workflow template rather than a blog. See unverified., resolution
+  1024x1024 in every BFL HF card example and in CivitAI's 9B training recipe
+  ("resolution": 1024). Non-square ratios work but no official list of
+  supported dimensions or a stated alignment multiple was found — see
+  unverified., lora_strength 1.0 is the diffusers/ComfyUI default and what
+  CivitAI applies to training previews (strength 1.0). Published per-LoRA
+  guidance varies and should be read from the LoRA, not assumed: the widely
+  used Klein 9B consistency LoRA's CivitAI article says start at 0.5 within a
+  0.2-1.0 range (0.2-0.4 loose, 0.5-0.7 balanced, 0.8-1.0 maximum structural
+  stability) and warns that too high a strength stops edits from applying;
+  SOLRICKS' Realistic Detail LoRA card says 0.6-0.8 subtle, 0.8-1.0 stronger,
+  start 0.8; the vafipas663 distillation-delta LoRA card warns it "produces
+  static noise at strength > 0.5. And sometimes, you have to go as low as
+  0.1.".
+
+───────────────────────────────────────────────────────────────────────────────
+  ADOPTION & POPULARITY
+───────────────────────────────────────────────────────────────────────────────
+  Downloads               5,272
+  Rating                  100% positive (335/335)
+  Comments                0
+  Published               197 days ago
+  Download rate           2.4/day since release
+
+  Adoption rank           #91 of 150 in this collection  (Modest)
+  Momentum rank           #118 of 150 by download rate  (Long tail)
+
+  Adoption tier is the percentile of total downloads within this collection;
+  momentum is the percentile of downloads-per-day since release. Momentum is
+  ranked rather than measured against a fixed rate, because this base model
+  is itself new and every LoRA looks fast on an absolute scale. Both numbers
+  are CivitAI's, read on the scrape date at the foot of this file.
+
+───────────────────────────────────────────────────────────────────────────────
+  CHAINING / STACKING
+───────────────────────────────────────────────────────────────────────────────
+  Stacks with others      not stated
+
+  The author gave no chaining guidance. The rules below are the
+  general FLUX.2 [klein] 9B ones, not anything specific to this LoRA.
+
+  General FLUX.2 [klein] 9B rules that apply here (see REFERENCE.txt):
+
+  • Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model
+    output into model input. There is no CLIP side to route — klein LoRAs
+    are model-only because the Qwen3 text encoder is not trained (CivitAI's
+    recipe fixes trainTextEncoder: false).
+  • Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B
+    LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden,
+    5+20 vs 8+24 blocks), not a quality question.
+  • Mixing a "Flux.2 Klein 9B" LoRA and a "Flux.2 Klein 9B-base" LoRA in
+    one stack is fine — those strings describe the training checkpoint, not
+    incompatible tensor layouts. Set the sampler for the checkpoint
+    actually loaded.
+  • Run base-trained LoRAs on the distilled checkpoint when you want speed:
+    HF's klein LoRA guide reports the adapter loads there and typically
+    gives better results than on base, at 4 steps.
+  • A base-to-turbo / distillation-extraction LoRA is a whole-model delta,
+    not a content LoRA. Put it on the base checkpoint only, and treat its
+    strength as a steps/CFG dial rather than a style knob. Do not add it on
+    top of the already-distilled checkpoint.
+  • Read each LoRA's own published strength range instead of defaulting
+    everything to 1.0 — klein LoRAs in the wild ship very different ones
+    (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA,
+    under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).
+  • With a consistency or edit LoRA in the stack, excessive strength
+    suppresses the prompt: the consistency LoRA's article says that if
+    edits stop applying, lower it before changing anything else.
+  • Prefer runtime LoRA patching over baking a LoRA into the checkpoint:
+    ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein
+    models (4B or 9B) since a March 2026 update — issue #13637, still open.
+  • The author never states which weight family this targets. Two different
+    questions, two different answers. Within 9B (distilled vs base — the
+    two CivitAI strings): not a real incompatibility. Both load through the
+    same Flux2KleinPipeline and have the same repo layout; for the 4B pair,
+    where both transformer configs are public, the distil
+
+───────────────────────────────────────────────────────────────────────────────
+  PROS
+───────────────────────────────────────────────────────────────────────────────
+  + Widely adopted — 5,272 downloads. [source: CivitAI stats]
+  + Positively received — 335 thumbs-up with no down-votes. [source:
+    CivitAI stats]
+  + Explicit trigger word(s) — 3DRenderAF, 3D Render Style — so the effect
+    can be turned on and off from the prompt. [source: model metadata]
+  + Compact at 158 MB — cheap to stack with other LoRAs. [source: file
+    size]
+
+───────────────────────────────────────────────────────────────────────────────
+  CONS & CAVEATS
+───────────────────────────────────────────────────────────────────────────────
+  - Author never states which generation mode it was trained for; test
+    against your own workflow before relying on it. [source: description]
+
+───────────────────────────────────────────────────────────────────────────────
+  COMPATIBILITY WARNING
+───────────────────────────────────────────────────────────────────────────────
+  Two different questions, two different answers. Within 9B (distilled vs
+  base — the two CivitAI strings): not a real incompatibility. Both load
+  through the same Flux2KleinPipeline and have the same repo layout; for the
+  4B pair, where both transformer configs are public, the distilled and base
+  configs are byte-for-byte identical in every shape field, and the 9B pair
+  is built the same way. HF's own klein LoRA guide states the workflow
+  directly: "Distilled models are step-compressed for fast inference; you
+  train against the base checkpoint and the adapter still loads on the
+  distilled model afterward", and "Applying the LoRA on the distilled model
+  typically gives better results than the base model, and it's faster" (that
+  post is written against 4B, but the family structure is the same). So the
+  two CivitAI strings record which checkpoint the author trained or previewed
+  against, not a hard compatibility wall. The real risk is a silent settings
+  mismatch: load a base-trained LoRA on the distilled checkpoint while
+  keeping the author's 20-50 step / CFG ~4 preview settings and the output
+  degrades badly (distilled wants 4 steps / guidance 1.0), or the reverse.
+  Nothing errors — it just looks wrong. Across sizes (9B vs 4B): a genuine
+  architectural mismatch, from the published transformer configs — different
+  hidden width (4096 vs 3072), different block counts (8+24 vs 5+20) and a
+  different text-conditioning dim (12288 vs 7680, following Qwen3-8B vs
+  Qwen3-4B). A 4B LoRA cannot patch a 9B model: expect missing keys and shape
+  errors, not degraded output. The exact ComfyUI error text is unverified. KV
+  variant: same 9B size and same 4-step distillation, but a distinct
+  checkpoint with its own pipeline; whether standard klein 9B LoRAs patch it
+  cleanly is not documented — see unverified.
+
+───────────────────────────────────────────────────────────────────────────────
+  AUTHOR'S DESCRIPTION
+───────────────────────────────────────────────────────────────────────────────
+  💖 Support My Work & Explore More If this model is useful to you, your
+  support helps me keep publishing free models and open-source AI tools.
+  - Patreon: Support on Patreon
+
+  - Ko-fi: Buy me a Ko-fi
+
+  - Buy Me a Coffee: Buy me a Coffee
+
+  Discover more of my models, projects, and practical AI resources:
+  - Portfolio: artificialguy.com — models, projects, and experiments.
+
+  - Prompt library: FindGoodPrompt — curated prompts for image and language
+  models.
+
+  - AI tools directory: JustGoodAI — a curated directory of useful AI tools.
+
+  - Twitter/X: @artificialguybr — new models announced here first.
+
+  3D.Redmond is here! Introducing 3D.Redmond, the ultimate LORA for creating
+  stunning 3D Render Style Images!
+  I'm grateful for the GPU time from Redmond.AI that allowed me to make this
+  LORA! If you need GPU, then you need the great services from Redmond.AI .
+  Test all my Loras here for free and unlimited. Thanks, HF, for Inference
+  API!
+  It is based on SD XL 1.0 and fine-tuned on a large dataset .
+  The LORA has a high capacity to generate 3D Render Style Images.
+  The tag for the model:3D Render Style, 3DRenderAF I really hope you like
+  the LORA and use it.
+
+──────────────���────────────────────────────────────────────────────────────────
+  EXAMPLE MEDIA
+───────────────────────────────────────────────────────────────────────────────
+  6 preview(s) in ./example_images/ — community results for
+  this LoRA. Videos keep a matching _poster.jpg still frame.
+
+  No prompts are recorded below: CivitAI's public API flags these posts as
+  having generation metadata but does not return it, so the prompts cannot be
+  scraped. Open the model page to read them.
+
+  01_image.jpeg  (image, 832x1216)
+  02_image.jpeg  (image, 832x1216)
+  03_image.jpeg  (image, 832x1216)
+  04_image.jpeg  (image, 832x1216)
+  05_image.jpeg  (image, 832x1216)
+  06_image.jpeg  (image, 832x1216)
+
+───────────────────────────────────────────────────────────────────────────────
+  HOW TO USE (COMFYUI)
+───────────────────────────────────────────────────────────────────────────────
+  1. Diffusion model -> ComfyUI/models/diffusion_models/. Distilled:
+  flux-2-klein-9b-fp8.safetensors. Base: flux-2-klein-
+  base-9b-fp8.safetensors. (BF16 single-file weights
+  flux-2-klein-9b.safetensors / flux-2-klein-base-9b.safetensors from the
+  gated HF repos also work if you have ~29GB; community GGUF builds exist
+  from unsloth and leejet for lower-VRAM setups.)
+  2. Text encoder ->
+  ComfyUI/models/text_encoders/qwen_3_8b_fp8mixed.safetensors (from Comfy-
+  Org/vae-text-encorder-for-flux-klein-9b). This is the 9B family's encoder;
+  the 4B family uses qwen_3_4b.safetensors — do not mix them.
+  3. VAE -> ComfyUI/models/vae/flux2-vae.safetensors (shared across klein
+  sizes).
+  4. LoRA -> ComfyUI/models/loras/ as a plain .safetensors.
+  5. Load with Load Diffusion Model (UNETLoader) for the transformer,
+  CLIPLoader for the Qwen3 text encoder, VAELoader for the VAE.
+  docs.comfy.org ships prebuilt Flux.2 klein graphs under Workflow Templates
+  for both the distilled and base branches — start from those rather than
+  wiring by hand, since neither the docs nor BFL's repo state
+  sampler/scheduler/shift in prose; those values live in the template.
+  6. Insert LoraLoaderModelOnly between the diffusion-model loader and the
+  sampler. Model-only is correct for klein: the Qwen3 text encoder is not
+  trained by the standard klein trainers, so there is no text-encoder half to
+  patch.
+  7. Set steps and guidance to match the checkpoint you actually loaded, not
+  the one the LoRA was trained on: 4 steps / guidance 1.0 on distilled,
+  ~20-50 steps / guidance ~4.0 on base.
+  8. Note the 9B repos are gated on HuggingFace — you must accept the FLUX
+  Non-Commercial License on the repo page before any download, including
+  scripted ones.
+
+───────────────────────────────────────────────────────────────────────────────
+  Scraped by lorakit on 2026-09-05 01:51 UTC.
+  Stats and description are the author's; pros/cons are derived from the
+  evidence tagged beside each line.
+───────────────────────────────────────────────────────────────────────────────

3d-redmond-3d-render-style-for-sd-xl/metadata.json

@@ -0,0 +1,135 @@
+{
+  "name": "3D.Redmond - 3D Render Style for SD XL",
+  "slug": "3d-redmond-3d-render-style-for-sd-xl",
+  "base_model": "FLUX.2 [klein] 9B",
+  "media_type": "image",
+  "model_id": 194138,
+  "version_id": 2705473,
+  "version_name": "v1.0 - Flux klein 9b",
+  "source": "civitai",
+  "url": "https://civitai.com/models/194138",
+  "repo_id": null,
+  "file_path": null,
+  "author": "artificialguybr",
+  "category": {
+    "key": "style",
+    "label": "Visual style"
+  },
+  "modalities": [],
+  "primary_modality": "unspecified",
+  "checkpoint_family": "unknown",
+  "trigger_words": [
+    "3DRenderAF",
+    "3D Render Style"
+  ],
+  "recommended_settings": {},
+  "chaining": {
+    "author": {
+      "stacks": "not stated",
+      "notes": "",
+      "avoid": [],
+      "requires": [],
+      "order": "",
+      "strength": "",
+      "stated": false
+    },
+    "general_rules": [
+      "Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model output into model input. There is no CLIP side to route — klein LoRAs are model-only because the Qwen3 text encoder is not trained (CivitAI's recipe fixes trainTextEncoder: false).",
+      "Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden, 5+20 vs 8+24 blocks), not a quality question.",
+      "Mixing a \"Flux.2 Klein 9B\" LoRA and a \"Flux.2 Klein 9B-base\" LoRA in one stack is fine — those strings describe the training checkpoint, not incompatible tensor layouts. Set the sampler for the checkpoint actually loaded.",
+      "Run base-trained LoRAs on the distilled checkpoint when you want speed: HF's klein LoRA guide reports the adapter loads there and typically gives better results than on base, at 4 steps.",
+      "A base-to-turbo / distillation-extraction LoRA is a whole-model delta, not a content LoRA. Put it on the base checkpoint only, and treat its strength as a steps/CFG dial rather than a style knob. Do not add it on top of the already-distilled checkpoint.",
+      "Read each LoRA's own published strength range instead of defaulting everything to 1.0 — klein LoRAs in the wild ship very different ones (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA, under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).",
+      "With a consistency or edit LoRA in the stack, excessive strength suppresses the prompt: the consistency LoRA's article says that if edits stop applying, lower it before changing anything else.",
+      "Prefer runtime LoRA patching over baking a LoRA into the checkpoint: ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein models (4B or 9B) since a March 2026 update — issue #13637, still open.",
+      "The author never states which weight family this targets. Two different questions, two different answers. Within 9B (distilled vs base — the two CivitAI strings): not a real incompatibility. Both load through the same Flux2KleinPipeline and have the same repo layout; for the 4B pair, where both transformer configs are public, the distil"
+    ]
+  },
+  "pros": [
+    "Widely adopted — 5,272 downloads. [source: CivitAI stats]",
+    "Positively received — 335 thumbs-up with no down-votes. [source: CivitAI stats]",
+    "Explicit trigger word(s) — 3DRenderAF, 3D Render Style — so the effect can be turned on and off from the prompt. [source: model metadata]",
+    "Compact at 158 MB — cheap to stack with other LoRAs. [source: file size]"
+  ],
+  "cons": [
+    "Author never states which generation mode it was trained for; test against your own workflow before relying on it. [source: description]"
+  ],
+  "file": {
+    "name": "[FLUX.2.Klein]3DRenderStyle_Redmond.safetensors",
+    "size_bytes": 165704504,
+    "size_mb": 158.03,
+    "sha256": "d364fc6eecd48d5c5eebb3f427b42596457903598005ae70ee73056de8a4fa83"
+  },
+  "stats": {
+    "downloads": 5272,
+    "thumbs_up": 335,
+    "thumbs_down": 0
+  },
+  "popularity": {
+    "downloads": 5272,
+    "thumbs_up": 335,
+    "thumbs_down": 0,
+    "comments": 0,
+    "approval": "100% positive (335/335)",
+    "age_days": 196.9,
+    "downloads_per_day": 2.35,
+    "rank_by_downloads": 91,
+    "rank_by_momentum": 118,
+    "of_total": 150,
+    "adoption_tier": "Modest",
+    "momentum": "Long tail"
+  },
+  "selection_score": 9692.0,
+  "selection_reasons": [
+    "5,272 downloads",
+    "335 thumbs-up",
+    "documented triggers + usage notes",
+    "top-ranked for its category"
+  ],
+  "example_media": [
+    {
+      "file": "01_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "02_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "03_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "04_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "05_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "06_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    }
+  ],
+  "tags": [
+    "3d render",
+    "styles",
+    "style",
+    "3d"
+  ],
+  "scraped_at": "2026-09-05T01:51:58+00:00",
+  "path": "flux2-klein-9b/3d-redmond-3d-render-style-for-sd-xl"
+}

ars-niji-style/Niji2Klein9b.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a7bd6f8fe1db6be4dd353d95c627b87df833aa7e11be4e3d3144d4dacfbf841e
+size 82866704

ars-niji-style/info.txt

@@ -0,0 +1,236 @@
+═══════════════════════════════════════════════════════════════════════════════
+  ARS NIJI STYLE
+═══════════════════════════════════════════════════════════════════════════════
+
+───────────────────────────────────────────────────────────────────────────────
+  IDENTITY
+───────────────────────────────────────────────────────────────────────────────
+  Base model        FLUX.2 [klein] 9B
+  Generation mode   not declared by the author
+  Checkpoint        not determined — verify before use
+  Category          Visual style
+  Author            ArsMachina
+  Source            https://civitai.com/models/729510
+  Provenance        CivitAI · model 729510 · version 2734450
+
+───────────────────────────────────────────────────────────────────────────────
+  WEIGHTS
+───────────────────────────────────────────────────────────────────────────────
+  File              Niji2Klein9b.safetensors
+  Version           Klein9B
+  Size              79.0 MB
+  Format            SafeTensor / unknown precision
+  SHA256            a7bd6f8fe1db6be4dd353d95c627b87df833aa7e11be4e3d3144d4dacfbf841e
+  Published         2026-04-08
+
+───────────────────────────────────────────────────────────────────────────────
+  TRIGGER WORDS
+───────────────────────────────────────────────────────────────────────────────
+  › "ArsNijiStyle,"
+
+───────────────────────────────────────────────────────────────────────────────
+  RECOMMENDED SETTINGS
+───────────────────────────────────────────────────────────────────────────────
+  The author did not publish specific settings.
+  No author settings were published. Community baseline: steps Distilled
+  FLUX.2-klein-9B: 4 (HF card example; it is step-distilled to exactly this,
+  and more steps do not refine further). Base FLUX.2-klein-base-9B: 50 in the
+  HF card example and in ComfyUI's benchmark; community practice runs lower,
+  with ~20 for production and 8-12 for fast iteration reported in the CivitAI
+  consistency-LoRA article., cfg Distilled: guidance_scale 1.0 (HF card). It
+  is guidance-distilled and its transformer config sets
+  guidance_embeds=false, so raising CFG degrades rather than improves. Base:
+  guidance_scale 4.0 (HF card); the CivitAI consistency-LoRA article
+  recommends 3.5-4.5 for most edits., sampler Not stated by BFL or by
+  docs.comfy.org, and the official flux2 GitHub repo documents no sampler,
+  scheduler or shift values either. Community consensus is euler with a
+  resolution-dependent shifted-Euler / flow-match schedule (ComfyUI's Flux2
+  workflows use a scheduler that interpolates shift between base_shift and
+  max_shift by output resolution). Start from ComfyUI's downloadable Flux.2
+  klein workflow template rather than a blog. See unverified., resolution
+  1024x1024 in every BFL HF card example and in CivitAI's 9B training recipe
+  ("resolution": 1024). Non-square ratios work but no official list of
+  supported dimensions or a stated alignment multiple was found — see
+  unverified., lora_strength 1.0 is the diffusers/ComfyUI default and what
+  CivitAI applies to training previews (strength 1.0). Published per-LoRA
+  guidance varies and should be read from the LoRA, not assumed: the widely
+  used Klein 9B consistency LoRA's CivitAI article says start at 0.5 within a
+  0.2-1.0 range (0.2-0.4 loose, 0.5-0.7 balanced, 0.8-1.0 maximum structural
+  stability) and warns that too high a strength stops edits from applying;
+  SOLRICKS' Realistic Detail LoRA card says 0.6-0.8 subtle, 0.8-1.0 stronger,
+  start 0.8; the vafipas663 distillation-delta LoRA card warns it "produces
+  static noise at strength > 0.5. And sometimes, you have to go as low as
+  0.1.".
+
+───────────────────────────────────────────────────────────────────────────────
+  ADOPTION & POPULARITY
+───────────────────────────────────────────────────────────────────────────────
+  Downloads               4,737
+  Rating                  100% positive (360/360)
+  Comments                3
+  Published               149 days ago
+  Download rate           4.4/day since release
+
+  Adoption rank           #99 of 150 in this collection  (Modest)
+  Momentum rank           #69 of 150 by download rate  (Steady)
+
+  Adoption tier is the percentile of total downloads within this collection;
+  momentum is the percentile of downloads-per-day since release. Momentum is
+  ranked rather than measured against a fixed rate, because this base model
+  is itself new and every LoRA looks fast on an absolute scale. Both numbers
+  are CivitAI's, read on the scrape date at the foot of this file.
+
+───────────────────────────────────────────────────────────────────────────────
+  CHAINING / STACKING
+───────────────────────────────────────────────────────────────────────────────
+  Stacks with others      not stated
+
+  The author gave no chaining guidance. The rules below are the
+  general FLUX.2 [klein] 9B ones, not anything specific to this LoRA.
+
+  General FLUX.2 [klein] 9B rules that apply here (see REFERENCE.txt):
+
+  • Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model
+    output into model input. There is no CLIP side to route — klein LoRAs
+    are model-only because the Qwen3 text encoder is not trained (CivitAI's
+    recipe fixes trainTextEncoder: false).
+  • Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B
+    LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden,
+    5+20 vs 8+24 blocks), not a quality question.
+  • Mixing a "Flux.2 Klein 9B" LoRA and a "Flux.2 Klein 9B-base" LoRA in
+    one stack is fine — those strings describe the training checkpoint, not
+    incompatible tensor layouts. Set the sampler for the checkpoint
+    actually loaded.
+  • Run base-trained LoRAs on the distilled checkpoint when you want speed:
+    HF's klein LoRA guide reports the adapter loads there and typically
+    gives better results than on base, at 4 steps.
+  • A base-to-turbo / distillation-extraction LoRA is a whole-model delta,
+    not a content LoRA. Put it on the base checkpoint only, and treat its
+    strength as a steps/CFG dial rather than a style knob. Do not add it on
+    top of the already-distilled checkpoint.
+  • Read each LoRA's own published strength range instead of defaulting
+    everything to 1.0 — klein LoRAs in the wild ship very different ones
+    (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA,
+    under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).
+  • With a consistency or edit LoRA in the stack, excessive strength
+    suppresses the prompt: the consistency LoRA's article says that if
+    edits stop applying, lower it before changing anything else.
+  • Prefer runtime LoRA patching over baking a LoRA into the checkpoint:
+    ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein
+    models (4B or 9B) since a March 2026 update — issue #13637, still open.
+  • The author never states which weight family this targets. Two different
+    questions, two different answers. Within 9B (distilled vs base — the
+    two CivitAI strings): not a real incompatibility. Both load through the
+    same Flux2KleinPipeline and have the same repo layout; for the 4B pair,
+    where both transformer configs are public, the distil
+
+───────────────────────────────────────────────────────────────────────────────
+  PROS
+───────────────────────────────────────────────────────────────────────────────
+  + Well established — 4,737 downloads. [source: CivitAI stats]
+  + Positively received — 360 thumbs-up with no down-votes. [source:
+    CivitAI stats]
+  + Explicit trigger word(s) — ArsNijiStyle, — so the effect can be turned
+    on and off from the prompt. [source: model metadata]
+  + Compact at 79 MB — cheap to stack with other LoRAs. [source: file size]
+
+───────────────────────────────────────────────────────────────────────────────
+  CONS & CAVEATS
+───────────────────────────────────────────────────────────────────────────────
+  - Author never states which generation mode it was trained for; test
+    against your own workflow before relying on it. [source: description]
+
+───────────────────────────────────────────────────────────────────────────────
+  COMPATIBILITY WARNING
+───────────────────────────────────────────────────────────────────────────────
+  Two different questions, two different answers. Within 9B (distilled vs
+  base — the two CivitAI strings): not a real incompatibility. Both load
+  through the same Flux2KleinPipeline and have the same repo layout; for the
+  4B pair, where both transformer configs are public, the distilled and base
+  configs are byte-for-byte identical in every shape field, and the 9B pair
+  is built the same way. HF's own klein LoRA guide states the workflow
+  directly: "Distilled models are step-compressed for fast inference; you
+  train against the base checkpoint and the adapter still loads on the
+  distilled model afterward", and "Applying the LoRA on the distilled model
+  typically gives better results than the base model, and it's faster" (that
+  post is written against 4B, but the family structure is the same). So the
+  two CivitAI strings record which checkpoint the author trained or previewed
+  against, not a hard compatibility wall. The real risk is a silent settings
+  mismatch: load a base-trained LoRA on the distilled checkpoint while
+  keeping the author's 20-50 step / CFG ~4 preview settings and the output
+  degrades badly (distilled wants 4 steps / guidance 1.0), or the reverse.
+  Nothing errors — it just looks wrong. Across sizes (9B vs 4B): a genuine
+  architectural mismatch, from the published transformer configs — different
+  hidden width (4096 vs 3072), different block counts (8+24 vs 5+20) and a
+  different text-conditioning dim (12288 vs 7680, following Qwen3-8B vs
+  Qwen3-4B). A 4B LoRA cannot patch a 9B model: expect missing keys and shape
+  errors, not degraded output. The exact ComfyUI error text is unverified. KV
+  variant: same 9B size and same 4-step distillation, but a distinct
+  checkpoint with its own pipeline; whether standard klein 9B LoRAs patch it
+  cleanly is not documented — see unverified.
+
+───────────────────────────────────────────────────────────────────────────────
+  AUTHOR'S DESCRIPTION
+───────────────────────────────────────────────────────────────────────────────
+  V2 Update: Ars Niji 2 - Now with 50% more Waifus Complete rework of the
+  dataset, so it reflects better the adjustments to my "Personal Style" in
+  the last year. This LoRA is trained on 500 handpicked Niji 6 images that
+  use my 'Personalized Style' The Trigger is ArsNijiStyle It works well in
+  the 0.4 – 1.5 range. Best results found between 0.8 – 1.2 , depending on
+  you LoRA stack.
+
+───────────────────────────────────────────────────────────────────────────────
+  EXAMPLE MEDIA
+───────────────────────────────────────────────────────────────────────────────
+  6 preview(s) in ./example_images/ — community results for
+  this LoRA. Videos keep a matching _poster.jpg still frame.
+
+  No prompts are recorded below: CivitAI's public API flags these posts as
+  having generation metadata but does not return it, so the prompts cannot be
+  scraped. Open the model page to read them.
+
+  01_image.jpeg  (image, 832x1216)
+  02_image.jpeg  (image, 1664x2944)
+  03_image.jpeg  (image, 1024x1024)
+  04_image.jpeg  (image, 832x1216)
+  05_image.jpeg  (image, 1216x832)
+  06_image.jpeg  (image, 1200x1920)
+
+───────────────────────────────────────────────────────────────────────────────
+  HOW TO USE (COMFYUI)
+───────────────────────────────────────────────────────────────────────────────
+  1. Diffusion model -> ComfyUI/models/diffusion_models/. Distilled:
+  flux-2-klein-9b-fp8.safetensors. Base: flux-2-klein-
+  base-9b-fp8.safetensors. (BF16 single-file weights
+  flux-2-klein-9b.safetensors / flux-2-klein-base-9b.safetensors from the
+  gated HF repos also work if you have ~29GB; community GGUF builds exist
+  from unsloth and leejet for lower-VRAM setups.)
+  2. Text encoder ->
+  ComfyUI/models/text_encoders/qwen_3_8b_fp8mixed.safetensors (from Comfy-
+  Org/vae-text-encorder-for-flux-klein-9b). This is the 9B family's encoder;
+  the 4B family uses qwen_3_4b.safetensors — do not mix them.
+  3. VAE -> ComfyUI/models/vae/flux2-vae.safetensors (shared across klein
+  sizes).
+  4. LoRA -> ComfyUI/models/loras/ as a plain .safetensors.
+  5. Load with Load Diffusion Model (UNETLoader) for the transformer,
+  CLIPLoader for the Qwen3 text encoder, VAELoader for the VAE.
+  docs.comfy.org ships prebuilt Flux.2 klein graphs under Workflow Templates
+  for both the distilled and base branches — start from those rather than
+  wiring by hand, since neither the docs nor BFL's repo state
+  sampler/scheduler/shift in prose; those values live in the template.
+  6. Insert LoraLoaderModelOnly between the diffusion-model loader and the
+  sampler. Model-only is correct for klein: the Qwen3 text encoder is not
+  trained by the standard klein trainers, so there is no text-encoder half to
+  patch.
+  7. Set steps and guidance to match the checkpoint you actually loaded, not
+  the one the LoRA was trained on: 4 steps / guidance 1.0 on distilled,
+  ~20-50 steps / guidance ~4.0 on base.
+  8. Note the 9B repos are gated on HuggingFace — you must accept the FLUX
+  Non-Commercial License on the repo page before any download, including
+  scripted ones.
+
+───────────────────────────────────────────────────────────────────────────────
+  Scraped by lorakit on 2026-09-05 01:52 UTC.
+  Stats and description are the author's; pros/cons are derived from the
+  evidence tagged beside each line.
+───────────────────────────────────────────────────────────────────────────────

ars-niji-style/metadata.json

@@ -0,0 +1,131 @@
+{
+  "name": "Ars Niji Style",
+  "slug": "ars-niji-style",
+  "base_model": "FLUX.2 [klein] 9B",
+  "media_type": "image",
+  "model_id": 729510,
+  "version_id": 2734450,
+  "version_name": "Klein9B",
+  "source": "civitai",
+  "url": "https://civitai.com/models/729510",
+  "repo_id": null,
+  "file_path": null,
+  "author": "ArsMachina",
+  "category": {
+    "key": "style",
+    "label": "Visual style"
+  },
+  "modalities": [],
+  "primary_modality": "unspecified",
+  "checkpoint_family": "unknown",
+  "trigger_words": [
+    "ArsNijiStyle,"
+  ],
+  "recommended_settings": {},
+  "chaining": {
+    "author": {
+      "stacks": "not stated",
+      "notes": "",
+      "avoid": [],
+      "requires": [],
+      "order": "",
+      "strength": "",
+      "stated": false
+    },
+    "general_rules": [
+      "Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model output into model input. There is no CLIP side to route — klein LoRAs are model-only because the Qwen3 text encoder is not trained (CivitAI's recipe fixes trainTextEncoder: false).",
+      "Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden, 5+20 vs 8+24 blocks), not a quality question.",
+      "Mixing a \"Flux.2 Klein 9B\" LoRA and a \"Flux.2 Klein 9B-base\" LoRA in one stack is fine — those strings describe the training checkpoint, not incompatible tensor layouts. Set the sampler for the checkpoint actually loaded.",
+      "Run base-trained LoRAs on the distilled checkpoint when you want speed: HF's klein LoRA guide reports the adapter loads there and typically gives better results than on base, at 4 steps.",
+      "A base-to-turbo / distillation-extraction LoRA is a whole-model delta, not a content LoRA. Put it on the base checkpoint only, and treat its strength as a steps/CFG dial rather than a style knob. Do not add it on top of the already-distilled checkpoint.",
+      "Read each LoRA's own published strength range instead of defaulting everything to 1.0 — klein LoRAs in the wild ship very different ones (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA, under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).",
+      "With a consistency or edit LoRA in the stack, excessive strength suppresses the prompt: the consistency LoRA's article says that if edits stop applying, lower it before changing anything else.",
+      "Prefer runtime LoRA patching over baking a LoRA into the checkpoint: ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein models (4B or 9B) since a March 2026 update — issue #13637, still open.",
+      "The author never states which weight family this targets. Two different questions, two different answers. Within 9B (distilled vs base — the two CivitAI strings): not a real incompatibility. Both load through the same Flux2KleinPipeline and have the same repo layout; for the 4B pair, where both transformer configs are public, the distil"
+    ]
+  },
+  "pros": [
+    "Well established — 4,737 downloads. [source: CivitAI stats]",
+    "Positively received — 360 thumbs-up with no down-votes. [source: CivitAI stats]",
+    "Explicit trigger word(s) — ArsNijiStyle, — so the effect can be turned on and off from the prompt. [source: model metadata]",
+    "Compact at 79 MB — cheap to stack with other LoRAs. [source: file size]"
+  ],
+  "cons": [
+    "Author never states which generation mode it was trained for; test against your own workflow before relying on it. [source: description]"
+  ],
+  "file": {
+    "name": "Niji2Klein9b.safetensors",
+    "size_bytes": 82866704,
+    "size_mb": 79.03,
+    "sha256": "a7bd6f8fe1db6be4dd353d95c627b87df833aa7e11be4e3d3144d4dacfbf841e"
+  },
+  "stats": {
+    "downloads": 4737,
+    "thumbs_up": 360,
+    "thumbs_down": 0
+  },
+  "popularity": {
+    "downloads": 4737,
+    "thumbs_up": 360,
+    "thumbs_down": 0,
+    "comments": 3,
+    "approval": "100% positive (360/360)",
+    "age_days": 149.4,
+    "downloads_per_day": 4.36,
+    "rank_by_downloads": 99,
+    "rank_by_momentum": 69,
+    "of_total": 150,
+    "adoption_tier": "Modest",
+    "momentum": "Steady"
+  },
+  "selection_score": 9457.0,
+  "selection_reasons": [
+    "4,737 downloads",
+    "360 thumbs-up",
+    "documented triggers + usage notes",
+    "top-ranked for its category"
+  ],
+  "example_media": [
+    {
+      "file": "01_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "02_image.jpeg",
+      "type": "image",
+      "resolution": "1664x2944",
+      "prompt": ""
+    },
+    {
+      "file": "03_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "04_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "05_image.jpeg",
+      "type": "image",
+      "resolution": "1216x832",
+      "prompt": ""
+    },
+    {
+      "file": "06_image.jpeg",
+      "type": "image",
+      "resolution": "1200x1920",
+      "prompt": ""
+    }
+  ],
+  "tags": [
+    "style"
+  ],
+  "scraped_at": "2026-09-05T01:52:02+00:00",
+  "path": "flux2-klein-9b/ars-niji-style"
+}

cutecartoon-redmond-cute-cartoon-lora/[FLUX.2.Klein]CuteCartoon_Redmond.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:77d80c4d6c0a377c9a44b979fd55376f63944dd2cdeedcc1587f840b4708dbc5
+size 165704504

cutecartoon-redmond-cute-cartoon-lora/info.txt

@@ -0,0 +1,257 @@
+═══════════════════════════════════════════════════════════════════════════════
+  CUTECARTOON.REDMOND - CUTE CARTOON LORA
+═══════════════════════════════════════════════════════════════════════════════
+
+───────────────────────────────────────────────────────────────────────────────
+  IDENTITY
+───────────────────────────────────────────────────────────────────────────────
+  Base model        FLUX.2 [klein] 9B
+  Generation mode   not declared by the author
+  Checkpoint        not determined — verify before use
+  Category          Visual style
+  Author            artificialguybr
+  Source            https://civitai.com/models/134720
+  Provenance        CivitAI · model 134720 · version 2717017
+
+───────────────────────────────────────────────────────────────────────────────
+  WEIGHTS
+───────────────────────────────────────────────────────────────────────────────
+  File              [FLUX.2.Klein]CuteCartoon_Redmond.safetensors
+  Version           v1.0 - Flux klein 9b
+  Size              158.0 MB
+  Format            SafeTensor / unknown precision
+  SHA256            77d80c4d6c0a377c9a44b979fd55376f63944dd2cdeedcc1587f840b4708dbc5
+  Published         2026-02-23
+
+───────────────────────────────────────────────────────────────────────────────
+  TRIGGER WORDS
+───────────────────────────────────────────────────────────────────────────────
+  › "CuteCartoonAF"
+  › "Cute Cartoon"
+
+───────────────────────────────────────────────────────────────────────────────
+  RECOMMENDED SETTINGS
+───────────────────────────────────────────────────────────────────────────────
+  The author did not publish specific settings.
+  No author settings were published. Community baseline: steps Distilled
+  FLUX.2-klein-9B: 4 (HF card example; it is step-distilled to exactly this,
+  and more steps do not refine further). Base FLUX.2-klein-base-9B: 50 in the
+  HF card example and in ComfyUI's benchmark; community practice runs lower,
+  with ~20 for production and 8-12 for fast iteration reported in the CivitAI
+  consistency-LoRA article., cfg Distilled: guidance_scale 1.0 (HF card). It
+  is guidance-distilled and its transformer config sets
+  guidance_embeds=false, so raising CFG degrades rather than improves. Base:
+  guidance_scale 4.0 (HF card); the CivitAI consistency-LoRA article
+  recommends 3.5-4.5 for most edits., sampler Not stated by BFL or by
+  docs.comfy.org, and the official flux2 GitHub repo documents no sampler,
+  scheduler or shift values either. Community consensus is euler with a
+  resolution-dependent shifted-Euler / flow-match schedule (ComfyUI's Flux2
+  workflows use a scheduler that interpolates shift between base_shift and
+  max_shift by output resolution). Start from ComfyUI's downloadable Flux.2
+  klein workflow template rather than a blog. See unverified., resolution
+  1024x1024 in every BFL HF card example and in CivitAI's 9B training recipe
+  ("resolution": 1024). Non-square ratios work but no official list of
+  supported dimensions or a stated alignment multiple was found — see
+  unverified., lora_strength 1.0 is the diffusers/ComfyUI default and what
+  CivitAI applies to training previews (strength 1.0). Published per-LoRA
+  guidance varies and should be read from the LoRA, not assumed: the widely
+  used Klein 9B consistency LoRA's CivitAI article says start at 0.5 within a
+  0.2-1.0 range (0.2-0.4 loose, 0.5-0.7 balanced, 0.8-1.0 maximum structural
+  stability) and warns that too high a strength stops edits from applying;
+  SOLRICKS' Realistic Detail LoRA card says 0.6-0.8 subtle, 0.8-1.0 stronger,
+  start 0.8; the vafipas663 distillation-delta LoRA card warns it "produces
+  static noise at strength > 0.5. And sometimes, you have to go as low as
+  0.1.".
+
+───────────────────────────────────────────────────────────────────────────────
+  ADOPTION & POPULARITY
+───────────────────────────────────────────────────────────────────────────────
+  Downloads               3,879
+  Rating                  100% positive (387/387)
+  Comments                1
+  Published               193 days ago
+  Download rate           1.5/day since release
+
+  Adoption rank           #116 of 150 in this collection  (Niche)
+  Momentum rank           #135 of 150 by download rate  (Long tail)
+
+  Adoption tier is the percentile of total downloads within this collection;
+  momentum is the percentile of downloads-per-day since release. Momentum is
+  ranked rather than measured against a fixed rate, because this base model
+  is itself new and every LoRA looks fast on an absolute scale. Both numbers
+  are CivitAI's, read on the scrape date at the foot of this file.
+
+───────────────────────────────────────────────────────────────────────────────
+  CHAINING / STACKING
+───────────────────────────────────────────────────────────────────────────────
+  Stacks with others      not stated
+
+  The author gave no chaining guidance. The rules below are the
+  general FLUX.2 [klein] 9B ones, not anything specific to this LoRA.
+
+  General FLUX.2 [klein] 9B rules that apply here (see REFERENCE.txt):
+
+  • Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model
+    output into model input. There is no CLIP side to route — klein LoRAs
+    are model-only because the Qwen3 text encoder is not trained (CivitAI's
+    recipe fixes trainTextEncoder: false).
+  • Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B
+    LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden,
+    5+20 vs 8+24 blocks), not a quality question.
+  • Mixing a "Flux.2 Klein 9B" LoRA and a "Flux.2 Klein 9B-base" LoRA in
+    one stack is fine — those strings describe the training checkpoint, not
+    incompatible tensor layouts. Set the sampler for the checkpoint
+    actually loaded.
+  • Run base-trained LoRAs on the distilled checkpoint when you want speed:
+    HF's klein LoRA guide reports the adapter loads there and typically
+    gives better results than on base, at 4 steps.
+  • A base-to-turbo / distillation-extraction LoRA is a whole-model delta,
+    not a content LoRA. Put it on the base checkpoint only, and treat its
+    strength as a steps/CFG dial rather than a style knob. Do not add it on
+    top of the already-distilled checkpoint.
+  • Read each LoRA's own published strength range instead of defaulting
+    everything to 1.0 — klein LoRAs in the wild ship very different ones
+    (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA,
+    under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).
+  • With a consistency or edit LoRA in the stack, excessive strength
+    suppresses the prompt: the consistency LoRA's article says that if
+    edits stop applying, lower it before changing anything else.
+  • Prefer runtime LoRA patching over baking a LoRA into the checkpoint:
+    ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein
+    models (4B or 9B) since a March 2026 update — issue #13637, still open.
+  • The author never states which weight family this targets. Two different
+    questions, two different answers. Within 9B (distilled vs base — the
+    two CivitAI strings): not a real incompatibility. Both load through the
+    same Flux2KleinPipeline and have the same repo layout; for the 4B pair,
+    where both transformer configs are public, the distil
+
+───────────────────────────────────────────────────────────────────────────────
+  PROS
+───────────────────────────────────────────────────────────────────────────────
+  + Well established — 3,879 downloads. [source: CivitAI stats]
+  + Positively received — 387 thumbs-up with no down-votes. [source:
+    CivitAI stats]
+  + Explicit trigger word(s) — CuteCartoonAF, Cute Cartoon — so the effect
+    can be turned on and off from the prompt. [source: model metadata]
+  + Compact at 158 MB — cheap to stack with other LoRAs. [source: file
+    size]
+
+───────────────────────────────────────────────────────────────────────────────
+  CONS & CAVEATS
+───────────────────────────────────────────────────────────────────────────────
+  - Author never states which generation mode it was trained for; test
+    against your own workflow before relying on it. [source: description]
+
+───────────────────────────────────────────────────────────────────────────────
+  COMPATIBILITY WARNING
+───────────────────────────────────────────────────────────────────────────────
+  Two different questions, two different answers. Within 9B (distilled vs
+  base — the two CivitAI strings): not a real incompatibility. Both load
+  through the same Flux2KleinPipeline and have the same repo layout; for the
+  4B pair, where both transformer configs are public, the distilled and base
+  configs are byte-for-byte identical in every shape field, and the 9B pair
+  is built the same way. HF's own klein LoRA guide states the workflow
+  directly: "Distilled models are step-compressed for fast inference; you
+  train against the base checkpoint and the adapter still loads on the
+  distilled model afterward", and "Applying the LoRA on the distilled model
+  typically gives better results than the base model, and it's faster" (that
+  post is written against 4B, but the family structure is the same). So the
+  two CivitAI strings record which checkpoint the author trained or previewed
+  against, not a hard compatibility wall. The real risk is a silent settings
+  mismatch: load a base-trained LoRA on the distilled checkpoint while
+  keeping the author's 20-50 step / CFG ~4 preview settings and the output
+  degrades badly (distilled wants 4 steps / guidance 1.0), or the reverse.
+  Nothing errors — it just looks wrong. Across sizes (9B vs 4B): a genuine
+  architectural mismatch, from the published transformer configs — different
+  hidden width (4096 vs 3072), different block counts (8+24 vs 5+20) and a
+  different text-conditioning dim (12288 vs 7680, following Qwen3-8B vs
+  Qwen3-4B). A 4B LoRA cannot patch a 9B model: expect missing keys and shape
+  errors, not degraded output. The exact ComfyUI error text is unverified. KV
+  variant: same 9B size and same 4-step distillation, but a distinct
+  checkpoint with its own pipeline; whether standard klein 9B LoRAs patch it
+  cleanly is not documented — see unverified.
+
+───────────────────────────────────────────────────────────────────────────────
+  AUTHOR'S DESCRIPTION
+───────────────────────────────────────────────────────────────────────────────
+  💖 Support My Work & Explore More If this model is useful to you, your
+  support helps me keep publishing free models and open-source AI tools.
+  - Patreon: Support on Patreon
+
+  - Ko-fi: Buy me a Ko-fi
+
+  - Buy Me a Coffee: Buy me a Coffee
+
+  Discover more of my models, projects, and practical AI resources:
+  - Portfolio: artificialguy.com — models, projects, and experiments.
+
+  - Prompt library: FindGoodPrompt — curated prompts for image and language
+  models.
+
+  - AI tools directory: JustGoodAI — a curated directory of useful AI tools.
+
+  - Twitter/X: @artificialguybr — new models announced here first.
+
+  CuteCartoon.Redmond is here! Introducing CuteCartoon.Redmond, the ultimate
+  LORA for creating funny cute images of characters!
+  I'm grateful for the GPU time from Redmond.AI that allowed me to make this
+  LORA! If you need GPU, then you need the great services from Redmond.AI .
+  The LORA has a high capacity to generate funny cute images of characters!
+  The tag for the model:CuteCartoonAF I really hope you like the LORA and use
+  it.
+
+───────────────────────────────────────────────────────────────────────────────
+  EXAMPLE MEDIA
+───────────────────────────────────────────────────────────────────────────────
+  6 preview(s) in ./example_images/ — community results for
+  this LoRA. Videos keep a matching _poster.jpg still frame.
+
+  No prompts are recorded below: CivitAI's public API flags these posts as
+  having generation metadata but does not return it, so the prompts cannot be
+  scraped. Open the model page to read them.
+
+  01_image.jpeg  (image, 1216x832)
+  02_image.jpeg  (image, 1024x1024)
+  03_image.jpeg  (image, 1024x1024)
+  04_image.jpeg  (image, 1024x1024)
+  05_image.jpeg  (image, 1216x832)
+  06_image.jpeg  (image, 1216x832)
+
+───────────────────────────────────────────────────────────────────────────────
+  HOW TO USE (COMFYUI)
+───────────────────────────────────────────────────────────────────────────────
+  1. Diffusion model -> ComfyUI/models/diffusion_models/. Distilled:
+  flux-2-klein-9b-fp8.safetensors. Base: flux-2-klein-
+  base-9b-fp8.safetensors. (BF16 single-file weights
+  flux-2-klein-9b.safetensors / flux-2-klein-base-9b.safetensors from the
+  gated HF repos also work if you have ~29GB; community GGUF builds exist
+  from unsloth and leejet for lower-VRAM setups.)
+  2. Text encoder ->
+  ComfyUI/models/text_encoders/qwen_3_8b_fp8mixed.safetensors (from Comfy-
+  Org/vae-text-encorder-for-flux-klein-9b). This is the 9B family's encoder;
+  the 4B family uses qwen_3_4b.safetensors — do not mix them.
+  3. VAE -> ComfyUI/models/vae/flux2-vae.safetensors (shared across klein
+  sizes).
+  4. LoRA -> ComfyUI/models/loras/ as a plain .safetensors.
+  5. Load with Load Diffusion Model (UNETLoader) for the transformer,
+  CLIPLoader for the Qwen3 text encoder, VAELoader for the VAE.
+  docs.comfy.org ships prebuilt Flux.2 klein graphs under Workflow Templates
+  for both the distilled and base branches — start from those rather than
+  wiring by hand, since neither the docs nor BFL's repo state
+  sampler/scheduler/shift in prose; those values live in the template.
+  6. Insert LoraLoaderModelOnly between the diffusion-model loader and the
+  sampler. Model-only is correct for klein: the Qwen3 text encoder is not
+  trained by the standard klein trainers, so there is no text-encoder half to
+  patch.
+  7. Set steps and guidance to match the checkpoint you actually loaded, not
+  the one the LoRA was trained on: 4 steps / guidance 1.0 on distilled,
+  ~20-50 steps / guidance ~4.0 on base.
+  8. Note the 9B repos are gated on HuggingFace — you must accept the FLUX
+  Non-Commercial License on the repo page before any download, including
+  scripted ones.
+
+───────────────────────────────────────────────────────────────────────────────
+  Scraped by lorakit on 2026-09-05 01:52 UTC.
+  Stats and description are the author's; pros/cons are derived from the
+  evidence tagged beside each line.
+───────────────────────────────────────────────────────────────────────────────

cutecartoon-redmond-cute-cartoon-lora/metadata.json

@@ -0,0 +1,134 @@
+{
+  "name": "CuteCartoon.Redmond - Cute Cartoon Lora",
+  "slug": "cutecartoon-redmond-cute-cartoon-lora",
+  "base_model": "FLUX.2 [klein] 9B",
+  "media_type": "image",
+  "model_id": 134720,
+  "version_id": 2717017,
+  "version_name": "v1.0 - Flux klein 9b",
+  "source": "civitai",
+  "url": "https://civitai.com/models/134720",
+  "repo_id": null,
+  "file_path": null,
+  "author": "artificialguybr",
+  "category": {
+    "key": "style",
+    "label": "Visual style"
+  },
+  "modalities": [],
+  "primary_modality": "unspecified",
+  "checkpoint_family": "unknown",
+  "trigger_words": [
+    "CuteCartoonAF",
+    "Cute Cartoon"
+  ],
+  "recommended_settings": {},
+  "chaining": {
+    "author": {
+      "stacks": "not stated",
+      "notes": "",
+      "avoid": [],
+      "requires": [],
+      "order": "",
+      "strength": "",
+      "stated": false
+    },
+    "general_rules": [
+      "Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model output into model input. There is no CLIP side to route — klein LoRAs are model-only because the Qwen3 text encoder is not trained (CivitAI's recipe fixes trainTextEncoder: false).",
+      "Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden, 5+20 vs 8+24 blocks), not a quality question.",
+      "Mixing a \"Flux.2 Klein 9B\" LoRA and a \"Flux.2 Klein 9B-base\" LoRA in one stack is fine — those strings describe the training checkpoint, not incompatible tensor layouts. Set the sampler for the checkpoint actually loaded.",
+      "Run base-trained LoRAs on the distilled checkpoint when you want speed: HF's klein LoRA guide reports the adapter loads there and typically gives better results than on base, at 4 steps.",
+      "A base-to-turbo / distillation-extraction LoRA is a whole-model delta, not a content LoRA. Put it on the base checkpoint only, and treat its strength as a steps/CFG dial rather than a style knob. Do not add it on top of the already-distilled checkpoint.",
+      "Read each LoRA's own published strength range instead of defaulting everything to 1.0 — klein LoRAs in the wild ship very different ones (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA, under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).",
+      "With a consistency or edit LoRA in the stack, excessive strength suppresses the prompt: the consistency LoRA's article says that if edits stop applying, lower it before changing anything else.",
+      "Prefer runtime LoRA patching over baking a LoRA into the checkpoint: ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein models (4B or 9B) since a March 2026 update — issue #13637, still open.",
+      "The author never states which weight family this targets. Two different questions, two different answers. Within 9B (distilled vs base — the two CivitAI strings): not a real incompatibility. Both load through the same Flux2KleinPipeline and have the same repo layout; for the 4B pair, where both transformer configs are public, the distil"
+    ]
+  },
+  "pros": [
+    "Well established — 3,879 downloads. [source: CivitAI stats]",
+    "Positively received — 387 thumbs-up with no down-votes. [source: CivitAI stats]",
+    "Explicit trigger word(s) — CuteCartoonAF, Cute Cartoon — so the effect can be turned on and off from the prompt. [source: model metadata]",
+    "Compact at 158 MB — cheap to stack with other LoRAs. [source: file size]"
+  ],
+  "cons": [
+    "Author never states which generation mode it was trained for; test against your own workflow before relying on it. [source: description]"
+  ],
+  "file": {
+    "name": "[FLUX.2.Klein]CuteCartoon_Redmond.safetensors",
+    "size_bytes": 165704504,
+    "size_mb": 158.03,
+    "sha256": "77d80c4d6c0a377c9a44b979fd55376f63944dd2cdeedcc1587f840b4708dbc5"
+  },
+  "stats": {
+    "downloads": 3879,
+    "thumbs_up": 387,
+    "thumbs_down": 0
+  },
+  "popularity": {
+    "downloads": 3879,
+    "thumbs_up": 387,
+    "thumbs_down": 0,
+    "comments": 1,
+    "approval": "100% positive (387/387)",
+    "age_days": 193.2,
+    "downloads_per_day": 1.55,
+    "rank_by_downloads": 116,
+    "rank_by_momentum": 135,
+    "of_total": 150,
+    "adoption_tier": "Niche",
+    "momentum": "Long tail"
+  },
+  "selection_score": 8923.0,
+  "selection_reasons": [
+    "3,879 downloads",
+    "387 thumbs-up",
+    "documented triggers + usage notes",
+    "top-ranked for its category"
+  ],
+  "example_media": [
+    {
+      "file": "01_image.jpeg",
+      "type": "image",
+      "resolution": "1216x832",
+      "prompt": ""
+    },
+    {
+      "file": "02_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "03_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "04_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "05_image.jpeg",
+      "type": "image",
+      "resolution": "1216x832",
+      "prompt": ""
+    },
+    {
+      "file": "06_image.jpeg",
+      "type": "image",
+      "resolution": "1216x832",
+      "prompt": ""
+    }
+  ],
+  "tags": [
+    "cartoon",
+    "cute",
+    "style"
+  ],
+  "scraped_at": "2026-09-05T01:52:19+00:00",
+  "path": "flux2-klein-9b/cutecartoon-redmond-cute-cartoon-lora"
+}

dark-fantasy-pulp-pinup/DarkFantasyPulpPinupKlein9b.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:691c6e3cb006088b2a9f8c7ba347d224b6b6d4cd972a3144b64782cca59e73d8
+size 82866736

dark-fantasy-pulp-pinup/info.txt

@@ -0,0 +1,235 @@
+═══════════════════════════════════════════════════════════════════════════════
+  DARK FANTASY PULP PINUP
+═══════════════════════════════════════════════════════════════════════════════
+
+───────────────────────────────────────────────────────────────────────────────
+  IDENTITY
+───────────────────────────────────────────────────────────────────────────────
+  Base model        FLUX.2 [klein] 9B
+  Generation mode   not declared by the author
+  Checkpoint        not determined — verify before use
+  Category          Visual style
+  Author            ArsMachina
+  Source            https://civitai.com/models/1550211
+  Provenance        CivitAI · model 1550211 · version 3159980
+
+───────────────────────────────────────────────────────────────────────────────
+  WEIGHTS
+───────────────────────────────────────────────────────────────────────────────
+  File              DarkFantasyPulpPinupKlein9b.safetensors
+  Version           Klein9B
+  Size              79.0 MB
+  Format            SafeTensor / unknown precision
+  SHA256            691c6e3cb006088b2a9f8c7ba347d224b6b6d4cd972a3144b64782cca59e73d8
+  Published         2026-08-14
+
+───────────────────────────────────────────────────────────────────────────────
+  TRIGGER WORDS
+───────────────────────────────────────────────────────────────────────────────
+  › "Dark Fantasy Pulp Pinup"
+
+───────────────────────────────────────────────────────────────────────────────
+  RECOMMENDED SETTINGS
+───────────────────────────────────────────────────────────────────────────────
+  The author did not publish specific settings.
+  No author settings were published. Community baseline: steps Distilled
+  FLUX.2-klein-9B: 4 (HF card example; it is step-distilled to exactly this,
+  and more steps do not refine further). Base FLUX.2-klein-base-9B: 50 in the
+  HF card example and in ComfyUI's benchmark; community practice runs lower,
+  with ~20 for production and 8-12 for fast iteration reported in the CivitAI
+  consistency-LoRA article., cfg Distilled: guidance_scale 1.0 (HF card). It
+  is guidance-distilled and its transformer config sets
+  guidance_embeds=false, so raising CFG degrades rather than improves. Base:
+  guidance_scale 4.0 (HF card); the CivitAI consistency-LoRA article
+  recommends 3.5-4.5 for most edits., sampler Not stated by BFL or by
+  docs.comfy.org, and the official flux2 GitHub repo documents no sampler,
+  scheduler or shift values either. Community consensus is euler with a
+  resolution-dependent shifted-Euler / flow-match schedule (ComfyUI's Flux2
+  workflows use a scheduler that interpolates shift between base_shift and
+  max_shift by output resolution). Start from ComfyUI's downloadable Flux.2
+  klein workflow template rather than a blog. See unverified., resolution
+  1024x1024 in every BFL HF card example and in CivitAI's 9B training recipe
+  ("resolution": 1024). Non-square ratios work but no official list of
+  supported dimensions or a stated alignment multiple was found — see
+  unverified., lora_strength 1.0 is the diffusers/ComfyUI default and what
+  CivitAI applies to training previews (strength 1.0). Published per-LoRA
+  guidance varies and should be read from the LoRA, not assumed: the widely
+  used Klein 9B consistency LoRA's CivitAI article says start at 0.5 within a
+  0.2-1.0 range (0.2-0.4 loose, 0.5-0.7 balanced, 0.8-1.0 maximum structural
+  stability) and warns that too high a strength stops edits from applying;
+  SOLRICKS' Realistic Detail LoRA card says 0.6-0.8 subtle, 0.8-1.0 stronger,
+  start 0.8; the vafipas663 distillation-delta LoRA card warns it "produces
+  static noise at strength > 0.5. And sometimes, you have to go as low as
+  0.1.".
+
+───────────────────────────────────────────────────────────────────────────────
+  ADOPTION & POPULARITY
+───────────────────────────────────────────────────────────────────────────────
+  Downloads               4,562
+  Rating                  100% positive (371/371)
+  Comments                0
+  Published               21 days ago
+  Download rate           4.8/day since release
+
+  Adoption rank           #103 of 150 in this collection  (Modest)
+  Momentum rank           #62 of 150 by download rate  (Steady · fresh release)
+
+  Adoption tier is the percentile of total downloads within this collection;
+  momentum is the percentile of downloads-per-day since release. Momentum is
+  ranked rather than measured against a fixed rate, because this base model
+  is itself new and every LoRA looks fast on an absolute scale. Both numbers
+  are CivitAI's, read on the scrape date at the foot of this file.
+
+───────────────────────────────────────────────────────────────────────────────
+  CHAINING / STACKING
+───────────────────────────────────────────────────────────────────────────────
+  Stacks with others      not stated
+
+  The author gave no chaining guidance. The rules below are the
+  general FLUX.2 [klein] 9B ones, not anything specific to this LoRA.
+
+  General FLUX.2 [klein] 9B rules that apply here (see REFERENCE.txt):
+
+  • Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model
+    output into model input. There is no CLIP side to route — klein LoRAs
+    are model-only because the Qwen3 text encoder is not trained (CivitAI's
+    recipe fixes trainTextEncoder: false).
+  • Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B
+    LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden,
+    5+20 vs 8+24 blocks), not a quality question.
+  • Mixing a "Flux.2 Klein 9B" LoRA and a "Flux.2 Klein 9B-base" LoRA in
+    one stack is fine — those strings describe the training checkpoint, not
+    incompatible tensor layouts. Set the sampler for the checkpoint
+    actually loaded.
+  • Run base-trained LoRAs on the distilled checkpoint when you want speed:
+    HF's klein LoRA guide reports the adapter loads there and typically
+    gives better results than on base, at 4 steps.
+  • A base-to-turbo / distillation-extraction LoRA is a whole-model delta,
+    not a content LoRA. Put it on the base checkpoint only, and treat its
+    strength as a steps/CFG dial rather than a style knob. Do not add it on
+    top of the already-distilled checkpoint.
+  • Read each LoRA's own published strength range instead of defaulting
+    everything to 1.0 — klein LoRAs in the wild ship very different ones
+    (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA,
+    under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).
+  • With a consistency or edit LoRA in the stack, excessive strength
+    suppresses the prompt: the consistency LoRA's article says that if
+    edits stop applying, lower it before changing anything else.
+  • Prefer runtime LoRA patching over baking a LoRA into the checkpoint:
+    ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein
+    models (4B or 9B) since a March 2026 update — issue #13637, still open.
+  • The author never states which weight family this targets. Two different
+    questions, two different answers. Within 9B (distilled vs base — the
+    two CivitAI strings): not a real incompatibility. Both load through the
+    same Flux2KleinPipeline and have the same repo layout; for the 4B pair,
+    where both transformer configs are public, the distil
+
+───────────────────────────────────────────────────────────────────────────────
+  PROS
+───────────────────────────────────────────────────────────────────────────────
+  + Well established — 4,562 downloads. [source: CivitAI stats]
+  + Positively received — 371 thumbs-up with no down-votes. [source:
+    CivitAI stats]
+  + Explicit trigger word(s) — Dark Fantasy Pulp Pinup — so the effect can
+    be turned on and off from the prompt. [source: model metadata]
+  + Compact at 79 MB — cheap to stack with other LoRAs. [source: file size]
+
+───────────────────────────────────────────────────────────────────────────────
+  CONS & CAVEATS
+───────────────────────────────────────────────────────────────────────────────
+  - Author never states which generation mode it was trained for; test
+    against your own workflow before relying on it. [source: description]
+
+───────────────────────────────────────────────────────────────────────────────
+  COMPATIBILITY WARNING
+───────────────────────────────────────────────────────────────────────────────
+  Two different questions, two different answers. Within 9B (distilled vs
+  base — the two CivitAI strings): not a real incompatibility. Both load
+  through the same Flux2KleinPipeline and have the same repo layout; for the
+  4B pair, where both transformer configs are public, the distilled and base
+  configs are byte-for-byte identical in every shape field, and the 9B pair
+  is built the same way. HF's own klein LoRA guide states the workflow
+  directly: "Distilled models are step-compressed for fast inference; you
+  train against the base checkpoint and the adapter still loads on the
+  distilled model afterward", and "Applying the LoRA on the distilled model
+  typically gives better results than the base model, and it's faster" (that
+  post is written against 4B, but the family structure is the same). So the
+  two CivitAI strings record which checkpoint the author trained or previewed
+  against, not a hard compatibility wall. The real risk is a silent settings
+  mismatch: load a base-trained LoRA on the distilled checkpoint while
+  keeping the author's 20-50 step / CFG ~4 preview settings and the output
+  degrades badly (distilled wants 4 steps / guidance 1.0), or the reverse.
+  Nothing errors — it just looks wrong. Across sizes (9B vs 4B): a genuine
+  architectural mismatch, from the published transformer configs — different
+  hidden width (4096 vs 3072), different block counts (8+24 vs 5+20) and a
+  different text-conditioning dim (12288 vs 7680, following Qwen3-8B vs
+  Qwen3-4B). A 4B LoRA cannot patch a 9B model: expect missing keys and shape
+  errors, not degraded output. The exact ComfyUI error text is unverified. KV
+  variant: same 9B size and same 4-step distillation, but a distinct
+  checkpoint with its own pipeline; whether standard klein 9B LoRAs patch it
+  cleanly is not documented — see unverified.
+
+───────────────────────────────────────────────────────────────────────────────
+  AUTHOR'S DESCRIPTION
+───────────────────────────────────────────────────────────────────────────────
+  Classic pin-up style infused with dark fantasy flair, this LoRA draws from
+  the energy of the 80s pulp art.
+  Expect striking figures, dramatic poses, and rich, moody atmospheres.
+  Inspired by the iconic works of Olivia De Berardinis , Luis Royo , and
+  Frank Frazetta .
+
+───────────────────────────────────────────────────────────────────────────────
+  EXAMPLE MEDIA
+───────────────────────────────────────────────────────────────────────────────
+  6 preview(s) in ./example_images/ — community results for
+  this LoRA. Videos keep a matching _poster.jpg still frame.
+
+  No prompts are recorded below: CivitAI's public API flags these posts as
+  having generation metadata but does not return it, so the prompts cannot be
+  scraped. Open the model page to read them.
+
+  01_image.jpeg  (image, 1024x1024)
+  02_image.jpeg  (image, 1024x1024)
+  03_image.jpeg  (image, 1024x1024)
+  04_image.jpeg  (image, 832x1216)
+  05_image.jpeg  (image, 832x1216)
+  06_image.jpeg  (image, 832x1216)
+
+───────────────────────────────────────────────────────────────────────────────
+  HOW TO USE (COMFYUI)
+───────────────────────────────────────────────────────────────────────────────
+  1. Diffusion model -> ComfyUI/models/diffusion_models/. Distilled:
+  flux-2-klein-9b-fp8.safetensors. Base: flux-2-klein-
+  base-9b-fp8.safetensors. (BF16 single-file weights
+  flux-2-klein-9b.safetensors / flux-2-klein-base-9b.safetensors from the
+  gated HF repos also work if you have ~29GB; community GGUF builds exist
+  from unsloth and leejet for lower-VRAM setups.)
+  2. Text encoder ->
+  ComfyUI/models/text_encoders/qwen_3_8b_fp8mixed.safetensors (from Comfy-
+  Org/vae-text-encorder-for-flux-klein-9b). This is the 9B family's encoder;
+  the 4B family uses qwen_3_4b.safetensors — do not mix them.
+  3. VAE -> ComfyUI/models/vae/flux2-vae.safetensors (shared across klein
+  sizes).
+  4. LoRA -> ComfyUI/models/loras/ as a plain .safetensors.
+  5. Load with Load Diffusion Model (UNETLoader) for the transformer,
+  CLIPLoader for the Qwen3 text encoder, VAELoader for the VAE.
+  docs.comfy.org ships prebuilt Flux.2 klein graphs under Workflow Templates
+  for both the distilled and base branches — start from those rather than
+  wiring by hand, since neither the docs nor BFL's repo state
+  sampler/scheduler/shift in prose; those values live in the template.
+  6. Insert LoraLoaderModelOnly between the diffusion-model loader and the
+  sampler. Model-only is correct for klein: the Qwen3 text encoder is not
+  trained by the standard klein trainers, so there is no text-encoder half to
+  patch.
+  7. Set steps and guidance to match the checkpoint you actually loaded, not
+  the one the LoRA was trained on: 4 steps / guidance 1.0 on distilled,
+  ~20-50 steps / guidance ~4.0 on base.
+  8. Note the 9B repos are gated on HuggingFace — you must accept the FLUX
+  Non-Commercial License on the repo page before any download, including
+  scripted ones.
+
+───────────────────────────────────────────────────────────────────────────────
+  Scraped by lorakit on 2026-09-05 01:52 UTC.
+  Stats and description are the author's; pros/cons are derived from the
+  evidence tagged beside each line.
+───────────────────────────────────────────────────────────────────────────────

dark-fantasy-pulp-pinup/metadata.json

@@ -0,0 +1,135 @@
+{
+  "name": "Dark Fantasy Pulp Pinup",
+  "slug": "dark-fantasy-pulp-pinup",
+  "base_model": "FLUX.2 [klein] 9B",
+  "media_type": "image",
+  "model_id": 1550211,
+  "version_id": 3159980,
+  "version_name": "Klein9B",
+  "source": "civitai",
+  "url": "https://civitai.com/models/1550211",
+  "repo_id": null,
+  "file_path": null,
+  "author": "ArsMachina",
+  "category": {
+    "key": "style",
+    "label": "Visual style"
+  },
+  "modalities": [],
+  "primary_modality": "unspecified",
+  "checkpoint_family": "unknown",
+  "trigger_words": [
+    "Dark Fantasy Pulp Pinup"
+  ],
+  "recommended_settings": {},
+  "chaining": {
+    "author": {
+      "stacks": "not stated",
+      "notes": "",
+      "avoid": [],
+      "requires": [],
+      "order": "",
+      "strength": "",
+      "stated": false
+    },
+    "general_rules": [
+      "Chain LoRAs by stringing LoraLoaderModelOnly nodes in series, model output into model input. There is no CLIP side to route — klein LoRAs are model-only because the Qwen3 text encoder is not trained (CivitAI's recipe fixes trainTextEncoder: false).",
+      "Only stack LoRAs of the same parameter size. 9B stacks with 9B; a 4B LoRA in a 9B stack is an architectural mismatch (3072 vs 4096 hidden, 5+20 vs 8+24 blocks), not a quality question.",
+      "Mixing a \"Flux.2 Klein 9B\" LoRA and a \"Flux.2 Klein 9B-base\" LoRA in one stack is fine — those strings describe the training checkpoint, not incompatible tensor layouts. Set the sampler for the checkpoint actually loaded.",
+      "Run base-trained LoRAs on the distilled checkpoint when you want speed: HF's klein LoRA guide reports the adapter loads there and typically gives better results than on base, at 4 steps.",
+      "A base-to-turbo / distillation-extraction LoRA is a whole-model delta, not a content LoRA. Put it on the base checkpoint only, and treat its strength as a steps/CFG dial rather than a style knob. Do not add it on top of the already-distilled checkpoint.",
+      "Read each LoRA's own published strength range instead of defaulting everything to 1.0 — klein LoRAs in the wild ship very different ones (0.5 start for the consistency LoRA, 0.8 for SOLRICKS' detail LoRA, under 0.5 and sometimes 0.1 for the vafipas663 distillation delta).",
+      "With a consistency or edit LoRA in the stack, excessive strength suppresses the prompt: the consistency LoRA's article says that if edits stop applying, lower it before changing anything else.",
+      "Prefer runtime LoRA patching over baking a LoRA into the checkpoint: ComfyUI's Model Merge / Model Save nodes have not detected Flux2 Klein models (4B or 9B) since a March 2026 update — issue #13637, still open.",
+      "The author never states which weight family this targets. Two different questions, two different answers. Within 9B (distilled vs base — the two CivitAI strings): not a real incompatibility. Both load through the same Flux2KleinPipeline and have the same repo layout; for the 4B pair, where both transformer configs are public, the distil"
+    ]
+  },
+  "pros": [
+    "Well established — 4,562 downloads. [source: CivitAI stats]",
+    "Positively received — 371 thumbs-up with no down-votes. [source: CivitAI stats]",
+    "Explicit trigger word(s) — Dark Fantasy Pulp Pinup — so the effect can be turned on and off from the prompt. [source: model metadata]",
+    "Compact at 79 MB — cheap to stack with other LoRAs. [source: file size]"
+  ],
+  "cons": [
+    "Author never states which generation mode it was trained for; test against your own workflow before relying on it. [source: description]"
+  ],
+  "file": {
+    "name": "DarkFantasyPulpPinupKlein9b.safetensors",
+    "size_bytes": 82866736,
+    "size_mb": 79.03,
+    "sha256": "691c6e3cb006088b2a9f8c7ba347d224b6b6d4cd972a3144b64782cca59e73d8"
+  },
+  "stats": {
+    "downloads": 4562,
+    "thumbs_up": 371,
+    "thumbs_down": 0
+  },
+  "popularity": {
+    "downloads": 4562,
+    "thumbs_up": 371,
+    "thumbs_down": 0,
+    "comments": 0,
+    "approval": "100% positive (371/371)",
+    "age_days": 21.5,
+    "downloads_per_day": 4.85,
+    "rank_by_downloads": 103,
+    "rank_by_momentum": 62,
+    "of_total": 150,
+    "adoption_tier": "Modest",
+    "momentum": "Steady · fresh release"
+  },
+  "selection_score": 9014.0,
+  "selection_reasons": [
+    "4,562 downloads",
+    "371 thumbs-up",
+    "top-ranked for its category"
+  ],
+  "example_media": [
+    {
+      "file": "01_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "02_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "03_image.jpeg",
+      "type": "image",
+      "resolution": "1024x1024",
+      "prompt": ""
+    },
+    {
+      "file": "04_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "05_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    },
+    {
+      "file": "06_image.jpeg",
+      "type": "image",
+      "resolution": "832x1216",
+      "prompt": ""
+    }
+  ],
+  "tags": [
+    "style",
+    "olivia de berardinis",
+    "dark fantasy",
+    "frank frazetta",
+    "pin-up",
+    "luis royo"
+  ],
+  "scraped_at": "2026-09-05T01:52:14+00:00",
+  "path": "flux2-klein-9b/dark-fantasy-pulp-pinup"
+}

delicate-whimsy/DelicateWhimsyKlein9b_000002000.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da19e77ed78b76874ddd1e1dcff895e2893d8c5d8179b004ab5c0f3a28c288fd
+size 82866720