Model
Trends
.ai
Model
Trends
.ai
os melhores modelos de IA de código aberto
Executar localmente
discm_anima_0.1.safetensors
132 MB · SafeTensor
Os downloads são servidos pelo Civitai. Alguns criadores pedem login no Civitai primeiro.
AD
<discm-anima>
Sobre este modelo
ANIMA · LoRAs
Criado por
llliii
no Civitai
by llliii - discm|Color Variations|brushstroke|Pattern|
Notas do criador
This LoRA was primarily trained on images with resolutions ranging from 1024 to 1400. Its key features include enhancing under-layer coloring (hidden colors), subsurface scattering (SSS), brushstroke textures, and intricate patterns. Because it covers a broad spectrum rather than a single element, relying solely on the trigger word will only bring out some common characteristics shared among the dataset. To unlock more diverse effects, you will need to utilize descriptive prompting. However, these combinations are difficult to list exhaustively, as mixing them can yield vastly different reactions. Therefore, the actual effects are demonstrated through the sample images. Some of these samples also utilize modified, high-quality prompts from Civitai. For generation settings: - To achieve better brushstroke textures (Recommended): Steps: 30 (Required) | Sampler: er_sde | Scheduler: bong_tangent - To retain the color effects while minimizing brushstrokes and excessive color blending: Steps: 24 | Sampler: er_sde | Scheduler: sgm_uniform Using the trigger word on its own won't fully showcase the LoRA's potential, but it minimizes disruption to the base model's core contents (such as specific characters or outfits). If you have any thoughts or feedback, please feel free to leave a comment! ---------------------------------------------------------------------------- このLoRAは、主に1024〜1400解像度の画像を使用して学習されています。 主な強化ポイントは、含み色（隠し色）、サブサーフェス・スキャッタリング（次表面散乱/SSS）、筆跡（ブラシアート）のテクスチャ、そして模様・パターンです。カバーしている範囲が単一の要素ではないため、トリガーワードのみの使用では素材に共通する一部の特徴しか再現されません。より詳細な効果を引き出すには、関連するプロンプト（記述）を組み合わせる必要があります。ただし、これらは一々列挙するのが難しく、組み合わせによって全く異なる反応を示すため、実際の一連の効果はサンプル画像にて展示しています。なお、一部のサンプル画像には、Civitaiで公開されている優れたプロンプトを改変して使用しています。 生成設定について： - より優れた筆跡テクスチャを得たい場合（推奨）： Steps: 30（必須） | Sampler: er_sde | Scheduler: bong_tangent - 色彩効果は欲しいが、筆跡や過度なグラデーション（色の変化）を抑えたい場合： Steps: 24 | Sampler: er_sde | Scheduler: sgm_uniform 単一のトリガーワードだけでは効果を完全に引き出すことはできませんが、その分、ベースモデルが持つ本来の内容（キャラクターや衣装など）への影響（破綻）を最小限に抑えることができます。何かアイデアやご意見がありましたら、お気軽にコメントを残してください。 ---------------------------------------------------------------------------- 이 LoRA는 주로 1024 ~ 1400 해상도의 이미지를 사용하여 학습되었습니다. 주요 강화 요소는 밑색(숨은 색), 서브서프에스 스캐터링(지하표면 산란/SSS), 붓터치 질감, 그리고 문양입니다. 커버하는 범위가 단일 항목이 아니기 때문에, 트리거 워드만 사용하면 데이터셋의 공통적인 특성 중 일부만 발현됩니다. 더 다양한 효과를 얻으려면 관련 서술(프롬프트)을 함께 사용해야 합니다. 하지만 이러한 조합을 일일이 열거하기 어렵고, 서로 조합되었을 때 완전히 다른 반응을 보이기 때문에 구체적인 효과는 예시 이미지를 통해 확인하실 수 있도록 준비했습니다. 일부 예시 이미지는 Civitai 사이트의 우수한 프롬프트를 수정하여 사용했습니다. 생성 설정 가이드: - 더 좋은 붓터치 질감을 얻고 싶을 때 (권장): Steps: 30 (필수) | Sampler: er_sde | Scheduler: bong_tangent - 색감 효과는 가져가되, 붓터치와 과도한 색상 전이(그라데이션)를 피하고 싶을 때: Steps: 24 | Sampler: er_sde | Scheduler: sgm_uniform 단일 트리거 워드만으로는 효과를 온전히 다 발현하기 어렵지만, 대신 베이스 모델의 고유 콘텐츠(캐릭터나 의상 등)에 대한 왜곡이나 파괴를 최소화할 수 있습니다. 좋은 아이디어나 의견이 있으시면 언제든지 댓글로 남겨주세요! ---------------------------------------------------------------------------- 此LoRA模型主要使用解析度在1024到1400之間的影像進行訓練。 關鍵強化因素包括基礎顏色（隱藏顏色）、次表面散射（SSS）、筆觸紋理和圖案。由於涵蓋的範圍並非單一項目，僅使用觸發詞只能產生資料集的部分常見特徵。為了獲得更廣泛的效果，您必須結合使用相關的描述（提示）。然而，很難逐一列出所有組合，而且由於它們組合起來會產生完全不同的響應，我們準備了一些示例圖像，以便您可以驗證具體效果。部分範例圖像是透過修改Civitai網站上的優秀提示而創建的。 產生設定指南： 當您想要獲得更好的筆觸紋理時（建議）： 步驟：30（必要）| 取樣器：er_sde |調度器：bong_tangent 當您想保留色調效果但避免筆觸和過多的顏色過渡（漸變）時： 步驟：24 | 取樣器：er_sde | 調度器：sgm_uniform 雖然僅憑一個觸發詞很難完全發揮效果，但它可以最大限度地減少對基礎模型獨特內容（例如角色或服裝）的扭曲或破壞。如果您有任何好的想法或意見，請隨時在評論區留言！ ---------------------------------------------------------------------------- 此LoRA模型主要使用分辨率在1024到1400之间的图像进行训练。 关键强化因素包括基础颜色（隐藏颜色）、次表面散射（SSS）、笔触纹理和图案。由于涵盖的范围并非单一项目，仅使用触发词只能产生数据集的部分常见特征。为了获得更广泛的效果，您必须结合使用相关的描述（提示）。然而，很难逐一列出所有组合，而且由于它们组合起来会产生完全不同的响应，我们准备了一些示例图像，以便您验证具体效果。部分示例图像是通过修改Civitai网站上的优秀提示而创建的。 生成设置指南： 当您想要获得更好的笔触纹理时（推荐）： 步骤：30（必需）| 采样器：er_sde |调度器：bong_tangent 当您想保留色调效果但避免笔触和过多的颜色过渡（渐变）时： 步骤：24 | 采样器：er_sde | 调度器：sgm_uniform 虽然仅凭一个触发词很难完全发挥效果，但它可以最大限度地减少对基础模型独特内容（例如角色或服装）的扭曲或破坏。如果您有任何好的想法或意见，请随时在评论区留言！ ----------------------------------------------------------------------------
Palavras-gatilho
@discm
262 downloads · 29 curtidas no Civitai
SHA256
2A63B07D82DFDC6C320DC2746EA099CEF55D79A6B4C6CD6A0B6B8EE256526E51
ID de modelo ModelTrends.ai
#25424
Denunciar conteúdo
Renders de exemplo
Prompts e configurações compartilhados por quem fez estes renders no Civitai. Escolha um render para ver como foi feito.
Configurações
Passos
30
Guidance
5
Seed
9
Tamanho
1024x1536
Prompt
(masterpiece, best quality, amazing quality, very aesthetic, extremely detailed, very detailed, incredibly absurdres, newest, highres, score_9, score_8), photorealistic,@Nijireol,@wlop, @jimu,victoria, a young girl standing in a field of yellow flowers. She is wearing a white dress and has long blonde hair that is styled in loose waves. The girl is holding a bouquet of sunflowers in her hands and is looking down at them with a peaceful expression on her face. The sky is blue and the sun is shining brightly, creating a warm and dreamy atmosphere. The overall mood of the image is peaceful and serene.
Prompt negativo
score_4, score_5, score_6, recent, glitch, deformed, mutated, ugly, disfigured, sketch, poorly drawn, lowres, low detail, bad anatomy, bad proportions, deformed anatomy, deformed face, deformed eyes, anatomically incorrect hands, missing fingers, extra digits, fewer digits, bad eye, multiple fingers, blurry eyes, extra legs, conjoined, stubble, ugly face, ugly eyes, letterbox, deformed, distorted hands, deformed hands, deformed fingers, 6 fingers, 4 fingers,
Configurações
Sampler
DPM++ 2M
Passos
30
Guidance
5
Seed
384672990665879
Tamanho
1024x1536
Prompt
(masterpiece, best quality, amazing quality, very aesthetic, extremely detailed, very detailed, incredibly absurdres, newest, highres, score_9, score_8), photorealistic,@Nijireol,@wlop, @jimu,victoria, 1girl, solo, full body, long hair, very long hair, blonde hair, hair flower, red flower, flower hair ornament, looking back, from behind, standing, sleeveless dress, long dress, sundress, floral print, floral dress, white dress, flowing dress, wind, windblown hair, wind lift, barefoot, outdoors, seaside, beach, ocean, sea, waves, horizon, blue sky, clouds, summer, sunlight, backlighting, ancient architecture, greek architecture, marble columns, column, pillar, colonnade, temple, stone floor, scenic, wide shot, low angle, framed by columns, depth of field
Prompt negativo
score_4, score_5, score_6, recent, glitch, deformed, mutated, ugly, disfigured, sketch, poorly drawn, lowres, low detail, bad anatomy, bad proportions, deformed anatomy, deformed face, deformed eyes, anatomically incorrect hands, missing fingers, extra digits, fewer digits, bad eye, multiple fingers, blurry eyes, extra legs, conjoined, stubble, ugly face, ugly eyes, letterbox, deformed, distorted hands, deformed hands, deformed fingers, 6 fingers, 4 fingers,
Configurações
Passos
30
Guidance
5
Seed
1101856919876507
Tamanho
768x1344
Prompt
(masterpiece, best quality, amazing quality, very aesthetic, extremely detailed, very detailed, incredibly absurdres, newest, highres, score_9, score_8), photorealistic,@Nijireol,victoria, 1girl, solo, anima_crossed_pistol_pose, two pistols, The face is above the pistols, pistol is in the bottom right corner, full_body,long blonde hair, rabbit ears, black and white nun, gold pistol,Long shot, Hitchcock zoom, church background
Prompt negativo
score_4, score_5, score_6, recent, glitch, deformed, mutated, ugly, disfigured, sketch, poorly drawn, lowres, low detail, bad anatomy, bad proportions, deformed anatomy, deformed face, deformed eyes, anatomically incorrect hands, missing fingers, extra digits, fewer digits, bad eye, multiple fingers, blurry eyes, extra legs, conjoined, stubble, ugly face, ugly eyes, letterbox, deformed, distorted hands, deformed hands, deformed fingers, 6 fingers, 4 fingers, thighhigh, close-up
Configurações
Sampler
er_sde
Passos
30
Guidance
4
Seed
129053007992302
Tamanho
896x1152
Prompt
masterpiece,@discm,score_9, score_8, year 2024, newest, best quality, highres, painting, chiaroscuro, pale skin,Shinosawa Hiro,gakuen Idolmaster,blonde hair,yellow eyes,black hair clip,her body angled sharply to the left while her head tilts forward. She wears a deep black dress, revealing her collarbone and extending into a neckline.In the dim night,skin is covered with intricate symmetrical golden tattoos, including the astral moon and flowers and straight line a thick heavy,ethereal flowing linework, wispy textures, surreal character design,
Prompt negativo
low quality, worst quality, normal quality, score_1, score_2, score_3, score_4, score_5, jpeg artifacts,bad hand
Configurações
Sampler
er_sde
Passos
30
Guidance
4
Seed
221473506026323
Tamanho
896x1152
Prompt
masterpiece,@discm,score_9, score_8, year 2024, newest, best quality, highres, painting, chiaroscuro, @96yottea, colored eyelashes, 1girl, Dark blue and Purple long gradient colors multicolored pink purple streaked hair, blue eyes, hair between eyes, multicolored eyes, rainbow eyes,Thick-rimmed glasses,pointing to the lips with the index finger, jewelry, pearl earrings, looking at viewer, open mouth, parted lips, solo, upper body, colorful, glowing, light particles, single earring, star shaped pupils, extremely detailed eyes, extreme closeup, eye focus, dark background,Neon lights,,
Prompt negativo
low quality, worst quality, normal quality, score_1, score_2, score_3, score_4, score_5, jpeg artifacts,bad hand
Configurações
Passos
30
Guidance
5
Tamanho
768x1344
Prompt
(masterpiece, best quality, amazing quality, very aesthetic, extremely detailed, very detailed, incredibly absurdres, newest, highres, score_9, score_8), photorealistic,@Nijireol,victoria, 1girl, solo, Gothic style, Little Red Riding Hood,rose elements, exquisite vintage longer rifle, gilded decorations
Prompt negativo
score_4, score_5, score_6, recent, glitch, deformed, mutated, ugly, disfigured, sketch, poorly drawn, lowres, low detail, bad anatomy, bad proportions, deformed anatomy, deformed face, deformed eyes, anatomically incorrect hands, missing fingers, extra digits, fewer digits, bad eye, multiple fingers, blurry eyes, extra legs, conjoined, stubble, ugly face, ugly eyes, letterbox, deformed, distorted hands, deformed hands, deformed fingers, 6 fingers, 4 fingers, thighhigh, close-up
Configurações
Sampler
er_sde_sgm_uniform
Passos
30
Guidance
4
Seed
273680750317787
Tamanho
1664x2432
Prompt
masterpiece, best quality, score_7, safe, 1girl, @discm. solo, dress, smiling.
Prompt negativo
worst quality, low quality, score_1, score_2, score_3. bad hands, huge breasts, loli, shiny skin. sepia, jpeg artifacts, artist name. head out of frame, monochrome, wet skin.
Configurações
Sampler
er_sde_simple
Passos
12
Guidance
1.5
Seed
675547631266234
Tamanho
1040x1520
Clip skip
2
Peso usado
1
Prompt
masterpiece, very aesthetic, @discm, painterly, brushstroke, impasto, oil painting \(medium\), impressionism, chiaroscuro, masterpiece, best quality, HDR, newest, score_9, highres, very aesthetic, high detail RAW color art, ultra detailed, score_8, 1girl, tashigi, one piece, blue eyes, glasses, red-framed eyewear, eyewear on head, black hair, short hair, long hair, blue hair, large breasts, katana, epaulettes, dynamic, jumping, 3d background, detailed background, outdoors, garden, flower pots, caustics
Prompt negativo
worst quality, low quality, score_1, score_2, score_3, blurry, jpeg artifacts, bad anatomy, bad hands, bad feet, missing fingers, extra fingers, text, watermark, logo, signature, dot patterns, sepia, censored, artist name, patreon logo, patreon username, twitter username, web address
Modelos de IA para geração de imagem, vídeo e texto
discm anima é um LoRA para a família Anima listado no ModelTrends.ai, um catálogo somente leitura de modelos de IA de código aberto para imagem, vídeo e texto. Compare-o com outros modelos Anima, veja a pontuação de calor para acompanhar a tendência e abra-o no PirateDiffusion ou BitVector para experimentar.
Navegar por família de modelos
Ver todos os modelos