Reprodução automática
Executar localmente
AstroFlatAnimeMH3-V01.safetensors
148 MB · SafeTensor · bf16
Os downloads são servidos pelo Civitai. Alguns criadores pedem login no Civitai primeiro.
Sobre este modelo
MINIMAX H3 · LoRAs
By Astroburner - (v1) Flat Vector / Game-Art Style LoRA for MiniMax H3. A style LoRA for MiniMax H3 focused on clean vector-inspired game art, polished flat illustration, modern anime-influenced character design, rich gradients and strong readable shapes. Try adding "Vector flat design illustration matching the reference style" into your description to activate this LoRA.
Notas do criador
# Flat Vector / Game-Art Style LoRA for MiniMax H3
A style LoRA for MiniMax H3 focused on clean vector-inspired game art, polished flat illustration, modern anime-influenced character design, rich gradients and strong readable shapes.
The training set was built from roughly 100 style-focused images covering characters, environments, romance scenes, action poses, fantasy, portraits, architecture and game/visual-novel-style compositions. The goal is not one single character or one fixed scene type, but a reusable visual language.
## Activation / Prompt Wording
This LoRA was trained with the original Krea-era caption wording, not with a newly invented MiniMax-H3-specific trigger token.
Use this early in style_and_tone:
Vector flat design illustration matching the reference style
For stronger style adherence, reinforce it with some of the recurring training language:
32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime style
You do not need to repeat every phrase every time. Start with the core activation wording, then add the descriptors that matter for the scene.
## How I Recommend Prompting It in MiniMax H3
Keep the visual style in style_and_tone.
Example:
style_and_tone: Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime style.
Then describe the actual scene separately in continuous_shot or multi_shot_sequence.
This helps keep the LoRA responsible for the visual identity while MiniMax H3 handles motion, camera movement, acting and timing.
## Good Things to Try
The LoRA is especially fun for:
- visual-novel and Ren'Py-style scenes
- romantic game cutscenes
- fantasy RPG party scenes
- JRPG combat and boss battles
- cyberpunk characters
- modern lifestyle scenes
- stylized commercials
- character portraits
- architecture and environment shots
- music-video concepts
- game trailers and animated trailer-style sequences
## MiniMax H3 Prompting Tips
For H3, I recommend separating the prompt into clear jobs:
- style_and_tone = visual look
- continuous_shot or multi_shot_sequence = action, camera and timing
- audio_design = music, ambience and sound effects
- write spoken dialogue explicitly when dialogue matters
For dialogue scenes, state who speaks and include the exact line.
For trailers, ads and game sequences, multi-shot timing works very well because each shot has a clear purpose.
For music-led videos, describe the music and its overall development, but avoid assigning sung lyric lines to strict time segments. Let the audio act as the master timeline while shot and camera cuts can still use timecodes.
If the result drifts too far toward realism, strengthen the style wording with:
clean vector artwork, smooth cel-shading, polished flat aesthetic, and professional flat design meets modern anime style.
If the result becomes too rigid or over-stylized, shorten the style section and let H3 handle more of the scene naturally.
## Examples
The attached example videos were generated with MiniMax H3 using this LoRA. I also included a separate prompt pack with 15 complete H3 Text-to-Video prompts covering film, games, romance, action, trailers, advertising and music-video scenarios.
Have fun with it, modify the prompts, and feel free to share your results. I am very curious to see what people build with the style.
Palavras-gatilho
see LoRA describtion
SHA256
1641FC424EB830FEA5CF0D54D305A7BF5BDC13C1E28D396D2CB36B9FE179D4B9
ID de modelo ModelTrends.ai
#27403
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
Sampler
Euler
Passos
25
Guidance
1
Tamanho
1376x768
Prompt
1. Target duration: 10 seconds
2. Format: 16:9
style_and_tone:
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime action style. Bold silhouettes, stylized smoke and debris, orange-red emergency lighting against cool blue shadows, high-energy graphic action.
audio_design:
Aggressive cinematic percussion and pulsing electronic bass. Heavy footsteps, distant gunfire, alarms, falling concrete, sparks and one strong explosion. Radio dialogue is slightly compressed but clearly intelligible.
continuous_shot:
[Continuous Shot] (00:00.000–10.000) Inside a damaged multi-level parking structure at dusk, an athletic female operative in dark tactical clothing runs across frame while the camera tracks beside her.
At 00:02.500 she dives behind a concrete pillar as debris strikes the floor nearby.
At 00:03.000 a male voice crackles through her radio:
"Exit's blocked. You've got thirty seconds."
She checks the corridor, reloads and replies:
"Then stop counting."
At 00:05.500 she bursts from cover and sprints toward a descending ramp. A stylized explosion erupts behind her, throwing bright orange fire and graphic debris across the background.
At 00:08.500 she lands after vaulting a barrier and says:
"Found another way."
The score hits a final heavy impact.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
25
Guidance
1
Tamanho
1376x768
Prompt
1. Target duration: 10 seconds
2. Format: 16:9
style_and_tone:
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime style. Cozy fantasy RPG party scene, warm candlelight, expressive ensemble characters, detailed but clean tavern environment.
audio_design:
Cheerful medieval tavern music with lute, flute and hand percussion. Background chatter, mugs touching, fireplace crackle and occasional laughter. Dialogue is lively and playful.
continuous_shot:
[Continuous Shot] (00:00.000–10.000) Four adventurers sit around a wooden table covered with mugs, maps and travel gear. The camera slowly glides around the group.
At 00:02.500 the female mage smiles at the warrior:
"You know the plan was to sneak inside."
The warrior raises his mug:
"We did sneak inside."
The rogue laughs:
"You kicked the gate off its hinges."
At 00:06.500 the warrior pauses and casually answers:
"Quietly."
The whole group bursts into laughter as warm candlelight glows across their faces.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
25
Guidance
1
Tamanho
1376x768
Prompt
1. Target duration: 10 seconds
2. Format: 16:9
style_and_tone:
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime cyberpunk style. Cyan-magenta neon palette, sleek geometric fashion, smoky club atmosphere, graphic reflections.
audio_design:
Heavy electronic club beat with deep stereo bass, muffled crowd noise, glasses, holographic interface sounds and distant machinery. Dialogue remains close and intelligible.
continuous_shot:
[Continuous Shot] (00:00.000–10.000) Inside a crowded futuristic nightclub, a female fixer waits beside a glowing bar while the camera moves through foreground dancers toward her.
At 00:03.000 she notices the protagonist and says:
"You're late."
A male voice replies:
"You said midnight."
She checks a holographic wrist display.
At 00:05.500 she says:
"That was before the target changed."
She gestures him closer and lowers her voice:
"Now we're stealing the whole building."
The bass drops as the camera ends on her confident half-smile.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
25
Guidance
1
Tamanho
1376x768
Prompt
1. Target duration: 12 seconds
2. Format: 16:9
style_and_tone:
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime trailer style. Stormy navy-blue palette, warm lighthouse highlights, bold graphic compositions, mystery-thriller atmosphere.
audio_design:
Begin with ocean wind and one low piano note. Gradually introduce dark strings, deep cinematic percussion and rising trailer pulses. Waves, thunder, creaking wood, lighthouse machinery, radio static and abrupt impact sounds. Deep restrained male trailer narrator.
multi_shot_sequence:
[Shot 1] (00:00.000–00:02.000) Aerial wide shot of a remote rocky island during a violent storm. An old lighthouse flashes through thick fog.
Narrator:
"Twenty years ago... the light went dark."
[Shot 2] (00:02.000–00:04.000) Inside the abandoned keeper's house, a woman in her early thirties enters carrying an old suitcase.
She whispers:
"Dad?"
[Shot 3] (00:04.000–00:06.000) Fast cuts: an aged family photograph, strange symbols in a journal, lighthouse gears suddenly turning, the woman's frightened eyes. Heavy trailer impact.
[Shot 4] (00:06.000–00:08.500) She runs up the spiral staircase while lights flicker.
A distorted radio voice says:
"Don't let it see you."
[Shot 5] (00:08.500–00:10.500) She reaches the lantern room. Far beyond the ocean, a mysterious geometric light answers the lighthouse beam.
She whispers:
"What is that?"
[Shot 6] (00:10.500–00:12.000) Hard cut to black. One lighthouse beam sweeps across frame.
Narrator:
"THE LAST LIGHTHOUSE."
Final deep impact.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
25
Guidance
1
Tamanho
1376x768
Prompt
1. Target duration: 10 seconds
2. Format: 16:9
style_and_tone:
ASTROFLATANIMEV01MH3, Vector flat design CG event illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, visual novel romantic story event CG, professional flat design meets modern anime VN storytelling style. Warm peach and lavender twilight, expressive romantic faces, elegant geometric hair.
audio_design:
Gentle piano and quiet acoustic guitar. Rooftop wind, distant traffic and soft city ambience. Female voice is shy and warm; unseen male protagonist voice is gentle and natural.
continuous_shot:
[Continuous Shot] (00:00.000–10.000) On a quiet rooftop at twilight, a young woman in a light summer dress stands near a chain-link fence with glowing city lights behind her.
At 00:03.000 she looks toward the unseen protagonist and says softly:
"I've been trying to tell you something all week."
The unseen young man answers:
"Then tell me now."
She brushes a strand of hair behind her ear and steps half a pace closer.
At 00:06.500 she says:
"I like it when it's just the two of us."
The piano becomes warmer as the camera settles on her vulnerable smile.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
25
Guidance
1
Tamanho
768x1376
Prompt
1. Target duration: 10 seconds
2. Format: 9:16 vertical
style_and_tone:
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime commercial style. Elegant beauty-ad palette, luminous stylized skin, refined product composition, soft peach-white gradients.
audio_design:
Minimal elegant electronic-pop instrumental with soft sparkling tones. Gentle water sounds, subtle glass handling and clean product-transition sounds. Warm confident female commercial voice-over.
multi_shot_sequence:
[Shot 1] (00:00.000–00:02.000) Vertical close-up of an elegant adult woman in a bright minimalist bathroom applying one drop of serum to her cheek.
Voice-over:
"Your morning starts with one drop."
[Shot 2] (00:02.000–00:04.000) Product beauty shot of a premium glass serum bottle rotating on pale stone with water droplets and clean highlights.
[Shot 3] (00:04.000–00:06.500) She presses the serum into her skin and turns toward warm morning light.
Voice-over:
"Lightweight. Fresh. Made to glow."
[Shot 4] (00:06.500–00:08.500) She walks confidently through a bright stylized city street.
[Shot 5] (00:08.500–00:10.000) Clean vertical hero shot of the bottle against an elegant gradient background.
Voice-over:
"Start bright."
Soft sonic logo.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
1
Guidance
1
Tamanho
768x1376
Prompt
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime cyberpunk style.
integrated_multimodal_description:
[Shot 1] Single continuous shot with no cuts. A spectacular high-altitude alpine base-jumping scene viewed entirely from a handheld selfie-stick action-camera perspective. A fit young woman in her late twenties stands on the narrow rocky edge of a very high mountain cliff in the European Alps. Vast snow-covered mountain peaks, steep rocky ridges, deep green valleys, and distant clouds stretch far behind her. The weather is clear and bright with strong natural daylight and excellent visibility.
She wears a sporty fitted base-jumping outfit with a premium extreme-sports aesthetic: predominantly deep navy blue with restrained red and warm yellow accent panels, modern aerodynamic fabric, lightweight gloves, reinforced boots, a visible parachute harness, and a compact parachute pack. The outfit feels professional and energetic without displaying any recognizable logos or branding.
She holds a long selfie stick in her right hand. The wide-angle camera faces back toward her, keeping her face, shoulders, torso, extended arm, and the enormous mountain drop clearly visible behind her. Slight natural action-camera lens distortion enhances the sense of height and speed. Strong alpine wind moves loose strands of her hair, pulls at her clothing, and creates subtle vibration in the selfie stick.
She looks directly into the camera with excited nervous energy, smiles broadly, raises her eyebrows and says in English, "Hey guys! This is gonna be insane. Follow me!"
Without any cut, she turns her upper body slightly toward the cliff edge while still keeping the selfie camera pointed back at herself. She takes a quick breath, shifts her weight, then takes two powerful running steps forward and launches herself cleanly from the cliff.
The camera remains continuously attached to the same selfie-stick perspective throughout the jump. The rock edge instantly disappears beneath her as she enters open freefall. The enormous alpine landscape rushes upward behind her while distant mountains remain visible across the horizon. Her expression changes from concentrated anticipation into exhilaration. Wind pressure becomes visibly stronger against her clothing and face.
She spreads her free arm outward, adjusts her body position, stabilizes naturally in the air, then looks back directly into the camera with a huge adrenaline-filled smile. During the final seconds she laughs breathlessly while continuing to fall through the vast alpine landscape. Motion remains smooth, physically convincing, and readable while preserving her appearance, outfit, camera position, and environment throughout the entire uninterrupted shot.
overall_soundscape:
Strong alpine wind surrounds the woman from the beginning, mixed with fabric flutter, harness movement, footsteps scraping against rock, and her breathing before the jump. Immediately after leaving the cliff, the sound expands into an intense continuous freefall air rush accompanied by stronger fabric vibration, breathy exertion, and brief natural laughter.
non_diegetic_music:
Fast electronic percussion with a tight rhythmic bass pulse begins underneath the opening moment. Short rising synth layers increase in intensity immediately after the jump while the rhythm remains energetic and controlled. No vocals.
Prompt negativo
n/a
Configurações
Sampler
Euler
Passos
20
Guidance
1
Tamanho
768x1376
Prompt
ASTROFLATANIMEV01MH3, Vector flat design illustration matching the reference style, 32-bit true color, rich gradient shading, premium flat design anime features, clean vector artwork, smooth cel-shading, polished flat aesthetic, professional flat design meets modern anime cyberpunk style.
integrated_multimodal_description:
[Shot 1] Single continuous vertical social-media shot with no cuts. Classic The Simpsons-inspired 2D cartoon character design blended carefully with the trained premium vector flat-design anime aesthetic. Preserve the recognizable clean black outlines, bright yellow skin, simplified facial shapes, large expressive eyes, iconic exaggerated proportions, bold flat colors, and instantly readable silhouette associated with the classic animated television style, while adding richer controlled cel-shading, smooth gradient accents, more refined clothing folds, and slightly greater anatomical definition.
Marge Simpson is instantly recognizable through her extremely tall blue beehive hairstyle, yellow skin, familiar eyes, nose, face shape, and overall cartoon identity. She appears as an adult woman with a more deliberately defined and athletic body than her ordinary television appearance while remaining clearly within the same stylized cartoon universe.
Her arms and legs have subtle visible muscle tone, particularly around the shoulders, upper arms, calves, and thighs. Her waist is clearly defined and her natural feminine silhouette is emphasized through the design of her clothing. Her proportions remain stylized rather than photorealistic.
Instead of her normal green dress, she wears a glamorous fitted black off-the-shoulder mini dress designed specifically for a modern nightlife look. The dress has a structured fitted waist, body-hugging silhouette, short but tasteful hemline, subtle side detailing, and smooth fabric that clearly follows her torso, hips, and legs. She wears simple black high heels, a small bracelet, and understated earrings. The outfit creates a confident sexy appearance without becoming explicit.
Marge stands full-body in the center of a bright stylized creator room designed as though she is recording a social-media dance video with a smartphone placed directly in front of her. The camera uses a fixed vertical 9:16 framing at approximately chest height, positioned far enough away to keep her entire body, including her feet and tall hairstyle, continuously visible.
She begins the dance by smiling toward the camera, shifting her weight rhythmically from one hip to the other while coordinating relaxed arm movements with the beat. Her facial expression is confident, playful, and slightly flirtatious.
She steps forward with one foot, moves both arms in a recognizable short-form dance gesture, then transitions into a smooth controlled hip movement synchronized with the rhythm. Her dress and body shapes remain consistent as she moves.
Without a cut, she rotates into a smooth half-turn, briefly presenting her profile and full silhouette to the camera. Her tall hair remains stable and recognizable while her arms move naturally with the rotation.
She turns back toward the camera, performs two final rhythmic steps, briefly places one hand near her hip, tilts her shoulders playfully, and finishes with a confident smiling pose directly toward the viewer.
A complete generic TikTok-style short-form social-media interface remains visible over the vertical video throughout the entire clip. It includes a circular profile icon area, username line, caption text area, heart icon with a visible like count, comment icon with a comment count, share icon, small music information area, and a rotating audio-disc graphic near the lower edge. The interface feels authentic to modern short-form social video while using fictional account information.
The social-media overlay remains stable and readable while Marge dances behind it. Her entire body stays clearly visible and centered for the majority of the performance, with no cinematic camera cuts or dramatic camera movement.
overall_soundscape:
Soft indoor room ambience, subtle heel contact with the floor, light clothing movement, and small natural movement sounds remain underneath the music. No spoken dialogue.
non_diegetic_music:
Catchy upbeat dance-pop instrumental with a strong electronic kick, bright synth chords, crisp claps, short melodic accents, and a bouncy bass groove designed for a viral short-form dance routine. The rhythm remains clear and consistent for the entire dance. No vocals.
Prompt negativo
n/a
Modelos semelhantes
Modelos de IA para geração de imagem, vídeo e texto
flat vector h3 é um LoRA para a família MiniMax H3 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 MiniMax H3, veja a pontuação de calor para acompanhar a tendência e abra-o no PirateDiffusion ou BitVector para experimentar.
Navegar por família de modelos
Modelos Text / LLM
411
Modelos Anima
509
Modelos Chroma
14
Modelos Flux
1,071
Modelos Flux 2 / Klein
1,237
Modelos MiniMax H3
331
Modelos Hunyuan
197
Modelos Ideogram
20
Modelos Krea2
674
Modelos Qwen
17
Modelos Qwen2
36
Modelos Ltx2
150
Modelos Stable Diffusion 1.5
6,465
Modelos Stable Diffusion XL 1.0
15,432
Modelos Zimage
804
Modelos Wan
919

Model
Trends
.ai
© ModelTrends.ai
|
Feito no Japão
|
© 2026








