Index/film-animation
House · filmmakers, animators, motion designers, storyboard artists
🎬

Filmmaking & Animation

100
Entries
4
Primitives

Showing 100 of 100 entries

session calloutsa-0

Krump Callout LoRA

Fine-tune Qwen/Qwen3-8B on 80 callout pairs so any LLM can host a real cypher.

SFT · Instruction
storyboard beatsa-1

Pixar Beat Sheet LoRA

Fine-tune Qwen/Qwen3-8B on 90 beat sheets so models draft three-act animation outlines.

SFT · Instruction
mouthshape timinga-2

Anime Lipflap LoRA

Fine-tune Qwen/Qwen3.5-4B on 60 phoneme-flap pairs so models suggest anime lip flaps.

SFT · Instruction
color grading rulesa-3

Noir Grade Rules LoRA

Fine-tune Qwen/Qwen3-8B on 70 scene-grade pairs so models emit noir grading notes.

SFT · Instruction
frame spacinga-4

Stopmo Frame LoRA

Fine-tune Qwen/Qwen3.5-4B on 50 motion-frame pairs so models advise stopmo increments.

SFT · Instruction
easing curvesa-5

Tween Ease LoRA

Fine-tune Qwen/Qwen3-8B on 85 ease-curve pairs so models name tween easings.

SFT · Instruction
character castinga-6

Casting Voice LoRA

Fine-tune Qwen/Qwen3-8B on 75 role-voice pairs so models propose animation voice casts.

SFT · Instruction
scene compositiona-7

Layout Grid LoRA

Fine-tune Qwen/Qwen3.5-4B on 65 grid-layout pairs so models suggest shot compositions.

SFT · Instruction
sound effect cuesa-8

Foley Cue LoRA

Fine-tune Qwen/Qwen3-8B on 80 action-foley pairs so models write foley cue sheets.

SFT · Instruction
shading stylea-9

Cel Shade LoRA

Fine-tune Qwen/Qwen3-8B on 70 light-shade pairs so models describe cel shading.

SFT · Instruction
rig hierarchy-10

Rig Naming LoRA

Fine-tune Qwen/Qwen3.5-4B on 55 char-rig pairs so models propose rig node names.

SFT · Instruction
shot slugging-11

Storyboard Slug LoRA

Fine-tune Qwen/Qwen3-8B on 90 script-slug pairs so models write board slugs.

SFT · Instruction
shot breakdown-12

VFX Breakdown LoRA

Fine-tune Qwen/Qwen3.5-9B on 85 shot-vfx pairs so models list VFX elements.

SFT · Instruction
motion smear-13

Cartoon Smear LoRA

Fine-tune Qwen/Qwen3.5-4B on 60 pose-smear pairs so models advise cartoon smears.

SFT · Instruction
puppet animation-14

Puppet Arm LoRA

Fine-tune Qwen/Qwen3-8B on 50 frame-arm pairs so models plan puppet moves.

SFT · Instruction
crowd cycles-15

Crowd Cycle LoRA

Fine-tune Qwen/Qwen3.5-9B on 75 crowd-walk pairs so models script crowd cycles.

SFT · Instruction
title typography-16

Title Kern LoRA

Fine-tune Qwen/Qwen3.5-4B on 65 title-kern pairs so models suggest title kerning.

SFT · Instruction
narrative structure-17

Act Structure LoRA

Fine-tune Qwen/Qwen3-8B on 88 premise-act pairs so models outline act structure.

SFT · Instruction
light setups-18

Light Rig LoRA

Fine-tune Qwen/Qwen3-8B on 70 mood-light pairs so models propose light rigs.

SFT · Instruction
deformation rules-19

Squash Stretch LoRA

Fine-tune Qwen/Qwen3.5-4B on 58 impact-squash pairs so models cue squash stretch.

SFT · Instruction
matte painting-20

Matte Paint LoRA

Fine-tune Qwen/Qwen3.5-9B on 72 scene-matte pairs so models brief matte paintings.

SFT · Instruction
dialogue timing-21

Dialogue Beat LoRA

Fine-tune Qwen/Qwen3-8B on 82 line-beat pairs so models time dialogue beats.

SFT · Instruction
prop animation-22

Prop Cycle LoRA

Fine-tune Qwen/Qwen3.5-4B on 53 prop-motion pairs so models plan prop cycles.

SFT · Instruction
render layers-23

Render Pass LoRA

Fine-tune Qwen/Qwen3.5-9B on 78 shot-pass pairs so models list render passes.

SFT · Instruction
facial acting-24

Expression Hook LoRA

Fine-tune Qwen/Qwen3-8B on 84 emotion-hook pairs so models suggest face hooks.

SFT · Instruction
music cue chartsr-0

Beat Chart Weaver

Fine-tune Qwen/Qwen3-8B on 60 (prompt, ChordPro) pairs so the model outputs a valid ChordPro sheet from a plain-language brief every time.

SFT · Structured Output
storyboard beatse-1

Storyboard Spine

Fine-tune Qwen/Qwen3.5-4B on 80 (synopsis, JSON shot list) pairs so storyboard artists get rigid shot schemas from scripts.

SFT · Structured Output
color grading rulesr-2

Color Grade Grammar

Fine-tune Qwen/Qwen3-8B on 70 (scene desc, LUT DSL) pairs so graders receive valid lookup-table scripts from notes.

SFT · Structured Output
character riggingr-3

Rig Node Schemer

Fine-tune Qwen/Qwen3.5-9B on 90 (char spec, node graph JSON) pairs so riggers get valid node trees from descriptions.

SFT · Structured Output
animation timingp-4

Frame Tempo Map

Fine-tune Qwen/Qwen3.5-4B on 50 (action, timing chart JSON) pairs so animators get beat-accurate frame charts.

SFT · Structured Output
scene layoute-5

Layout Grid Scribe

Fine-tune Qwen/Qwen3-8B on 75 (script line, layout YAML) pairs so layout artists receive valid blocking schemas.

SFT · Structured Output
dialogue syncn-6

Lip Sync Lexicon

Fine-tune Qwen/Qwen3.5-4B on 65 (line, viseme JSON) pairs so animators get rigid viseme timelines from speech.

SFT · Structured Output
virtual cinematographyr-7

Camera Move Coder

Fine-tune Qwen/Qwen3.5-9B on 85 (beat, camera DSL) pairs so operators get valid move scripts from story beats.

SFT · Structured Output
asset designr-8

Prop Spec Writer

Fine-tune Qwen/Qwen3-8B on 55 (concept, prop JSON) pairs so modelers receive valid asset schemas from notes.

SFT · Structured Output
effects setuph-9

FX Schema Smith

Fine-tune Qwen/Qwen3.5-9B on 95 (effect desc, Houdini JSON) pairs so TD's get valid FX node schemas.

SFT · Structured Output
shot planning-10

Shot List Loom

Fine-tune Qwen/Qwen3.5-4B on 45 (scene, shot CSV) pairs so directors get rigid shot tables from scenes.

SFT · Structured Output
character posing-11

Pose Library Link

Fine-tune Qwen/Qwen3-8B on 70 (emotion, pose JSON) pairs so animators get valid pose sets from briefs.

SFT · Structured Output
post editing-12

Edit Decision Drafter

Fine-tune Qwen/Qwen3.5-4B on 60 (assembly, EDL text) pairs so editors get valid edit lists from cuts.

SFT · Structured Output
production breakdown-13

Script Breakdown Bound

Fine-tune Qwen/Qwen3-8B on 80 (script page, breakdown JSON) pairs so producers get rigid element lists.

SFT · Structured Output
surface authoring-14

Texture Tag Weaver

Fine-tune Qwen/Qwen3.5-4B on 50 (surface, PBR JSON) pairs so texture artists get valid material schemas.

SFT · Structured Output
casting sheets-15

Voice Cast Coder

Fine-tune Qwen/Qwen3-8B on 65 (char, cast YAML) pairs so casting gets valid role schemas from bibles.

SFT · Structured Output
lighting setup-16

Light Rig Rune

Fine-tune Qwen/Qwen3.5-9B on 90 (mood, light JSON) pairs so lighters get valid rig schemas from notes.

SFT · Structured Output
compositing-17

Comp Node Knitter

Fine-tune Qwen/Qwen3.5-9B on 85 (shot, Nuke JSON) pairs so compositors get valid node trees from briefs.

SFT · Structured Output
procedural motion-18

Motion Curve Coder

Fine-tune Qwen/Qwen3.5-4B on 55 (move, curve DSL) pairs so TD's get valid easing scripts from intent.

SFT · Structured Output
graphic overlays-19

Title Card Templater

Fine-tune Qwen/Qwen3-8B on 60 (brief, title JSON) pairs so designers get valid overlay schemas from notes.

SFT · Structured Output
animatic assembly-20

Animatic Order Oracle

Fine-tune Qwen/Qwen3.5-4B on 45 (board set, animatic JSON) pairs so editors get valid sequence schemas.

SFT · Structured Output
production pipeline-21

Pipeline Task Tracer

Fine-tune Qwen/Qwen3.5-9B on 100 (task, DAG JSON) pairs so TDs get valid pipeline graphs from requests.

SFT · Structured Output
crowd animation-22

Crowd Sim Scriber

Fine-tune Qwen/Qwen3.5-9B on 80 (scene, crowd JSON) pairs so TDs get valid agent schemas from briefs.

SFT · Structured Output
render output-23

Render Pass Poet

Fine-tune Qwen/Qwen3-8B on 70 (shot, AOV JSON) pairs so lighters get valid pass lists from shots.

SFT · Structured Output
music spotting-24

Score Cue Cartographer

Fine-tune Qwen/Qwen3.5-4B on 50 (scene, cue JSON) pairs so music editors get valid spot schemas from reels.

SFT · Structured Output
narrative structurer-0

Beat Sheet Forger

Fine-tune Qwen/Qwen3-8B on 80 beat-sheet prompts so it emits three-act structures passing a parser-based coherence check.

RL · Programmatic Reward
cinematography notationh-1

Lens Grammar Smith

Fine-tune Qwen/Qwen3.5-4B on 60 shot-list specs so it outputs lens params passing a focal-length constraint grader.

RL · Programmatic Reward
character animationr-2

Pose Flow Coder

Fine-tune Qwen/Qwen3-8B on 70 keyframe texts so it generates joint angles passing inverse-kinematics feasibility checks.

RL · Programmatic Reward
color scriptingr-3

Color Script Tuner

Fine-tune Qwen/Qwen3.5-4B on 50 palette briefs so it returns hex ramps passing contrast-ratio grader tests.

RL · Programmatic Reward
sound design cuesr-4

Foley Labeler

Fine-tune Qwen/Qwen3-8B on 90 scene texts so it tags foley needs passing an ontology-match reward function.

RL · Programmatic Reward
storyboard beatsr-5

Storyboard Beat Miner

Fine-tune Qwen/Qwen3.5-9B on 100 script pages so it extracts panels passing a layout-density grader.

RL · Programmatic Reward
script polishingb-6

Dialogue Trim Lab

Fine-tune Qwen/Qwen3-8B on 80 dialogue cuts so it shortens lines passing a syllable-cap reward checker.

RL · Programmatic Reward
rigging rulese-7

Rig Constraint Sage

Fine-tune Qwen/Qwen3.5-4B on 60 rig specs so it emits constraints passing a cycle-free graph grader.

RL · Programmatic Reward
playback timingr-8

Frame Rate Planner

Fine-tune Qwen/Qwen3-8B on 70 timing briefs so it assigns fps passing a motion-blur budget grader.

RL · Programmatic Reward
asset breakdownr-9

Prop List Weaver

Fine-tune Qwen/Qwen3.5-4B on 50 scene scripts so it lists props passing a continuity-cross-check grader.

RL · Programmatic Reward
camera movement-10

Camera Move Theorist

Fine-tune Qwen/Qwen3-8B on 90 move prompts so it writes dolly paths passing a collision-avoidance grader.

RL · Programmatic Reward
phoneme timing-11

Lip Sync Writer

Fine-tune Qwen/Qwen3.5-9B on 100 line texts so it outputs phoneme spans passing a duration-match grader.

RL · Programmatic Reward
lighting setups-12

Lighting Key Crafter

Fine-tune Qwen/Qwen3-8B on 80 scene moods so it sets key lights passing a lux-balance grader.

RL · Programmatic Reward
edit transitions-13

Transition Logician

Fine-tune Qwen/Qwen3.5-4B on 60 edit lists so it picks cuts passing a rhythm-gap grader.

RL · Programmatic Reward
creature animation-14

Creature Gait Model

Fine-tune Qwen/Qwen3-8B on 70 gait specs so it describes cycles passing a foot-contact grader.

RL · Programmatic Reward
layout safety-15

Title Safe Architect

Fine-tune Qwen/Qwen3.5-4B on 50 frame comps so it places text passing a title-safe grader.

RL · Programmatic Reward
vfx shot breakdown-16

VFX Breakdown Scribe

Fine-tune Qwen/Qwen3-8B on 90 shot notes so it tags elements passing a pipeline-node grader.

RL · Programmatic Reward
dramatic pacing-17

Act Structure Auditor

Fine-tune Qwen/Qwen3.5-9B on 100 loglines so it flags acts passing a proportion-grader reward.

RL · Programmatic Reward
asset naming-18

Texture Naming Oracle

Fine-tune Qwen/Qwen3.5-4B on 60 asset dumps so it names maps passing a PBR-suffix grader.

RL · Programmatic Reward
animation curves-19

Motion Curve Shaper

Fine-tune Qwen/Qwen3-8B on 80 curve briefs so it outputs easing passing a jerk-limit grader.

RL · Programmatic Reward
runtime estimation-20

Script Page Timer

Fine-tune Qwen/Qwen3-8B on 70 pages so it predicts minutes passing a word-rate grader.

RL · Programmatic Reward
crowd placement-21

Crowd Layout Bot

Fine-tune Qwen/Qwen3.5-4B on 50 crowd briefs so it spaces agents passing a density-grader reward.

RL · Programmatic Reward
subtitle timing-22

Subtitle Splitter

Fine-tune Qwen/Qwen3-8B on 90 subtitle lines so it splits passing a CPS-grader reward.

RL · Programmatic Reward
puppet controls-23

Puppet Rig Poet

Fine-tune Qwen/Qwen3.5-9B on 100 control maps so it labels handles passing a hierarchy-grader reward.

RL · Programmatic Reward
sequence ordering-24

Scene Order Judge

Fine-tune Qwen/Qwen3-8B on 80 scene sets so it sequences passing a causality-grader reward.

RL · Programmatic Reward
studio house voicer-0

House Voice Taster

Fine-tune Qwen/Qwen3.5-4B on 40 preference triples so any script output lands in the studio's house voice.

RL · Preference Pairs
storyboard pacinge-1

Storyboard Beat Judge

Fine-tune Qwen/Qwen3-8B on 60 beat pairs so storyboard descriptions match pro pacing taste.

RL · Preference Pairs
color grading rulese-2

Color Grade Palate

Fine-tune Qwen/Qwen3.5-4B on 50 grading notes pairs so tone suggestions match senior grader taste.

RL · Preference Pairs
animatic timingr-3

Animatic Timing Ear

Fine-tune Qwen/Qwen3-8B on 70 timing pairs so animatic cut suggestions feel editorially correct.

RL · Preference Pairs
character rigginge-4

Rig Naming Sense

Fine-tune Qwen/Qwen3.5-4B on 45 naming pairs so rig node names follow studio convention taste.

RL · Preference Pairs
lip sync phrasinge-5

Lip Sync Cadence

Fine-tune Qwen/Qwen3-8B on 55 sync pairs so mouth-key descriptions match animator cadence taste.

RL · Preference Pairs
scene layoute-6

Layout Readability Eye

Fine-tune Qwen/Qwen3.5-4B on 65 layout pairs so layout notes match composition director taste.

RL · Preference Pairs
animation dialoguer-7

Dialogue Subtext Ear

Fine-tune Qwen/Qwen3-8B on 60 dialogue pairs so line rewrites keep intended subtext taste.

RL · Preference Pairs
FX simulatione-8

FX Chaos Taste

Fine-tune Qwen/Qwen3.5-9B on 75 FX pairs so sim parameter notes match lead FX artist taste.

RL · Preference Pairs
motion designl-9

Motion Curve Feel

Fine-tune Qwen/Qwen3.5-4B on 50 curve pairs so easing descriptions match designer motion taste.

RL · Preference Pairs
character casting-10

Casting Voice Match

Fine-tune Qwen/Qwen3-8B on 55 casting pairs so character descriptions match director casting taste.

RL · Preference Pairs
edit transitions-11

Edit Transition Gut

Fine-tune Qwen/Qwen3.5-4B on 45 transition pairs so transition ideas match editor instinct taste.

RL · Preference Pairs
prop design-12

Prop Design Sense

Fine-tune Qwen/Qwen3-8B on 65 prop pairs so prop briefs match art director visual taste.

RL · Preference Pairs
crowd animation-13

Crowd Shot Calm

Fine-tune Qwen/Qwen3.5-9B on 70 crowd pairs so crowd notes match supervisor density taste.

RL · Preference Pairs
title design-14

Title Card Rhythm

Fine-tune Qwen/Qwen3.5-4B on 40 title pairs so title card copy matches designer rhythm taste.

RL · Preference Pairs
compositing-15

Compositing Clean Eye

Fine-tune Qwen/Qwen3-8B on 60 comp pairs so comp notes match senior compositor clean taste.

RL · Preference Pairs
sound editorial-16

Sound Edit Hint

Fine-tune Qwen/Qwen3.5-4B on 50 sound pairs so sound cue notes match editor mic taste.

RL · Preference Pairs
puppet animation-17

Puppet Rig Logic

Fine-tune Qwen/Qwen3.5-4B on 45 puppet pairs so rig logic notes match puppeteer control taste.

RL · Preference Pairs
background painting-18

Background Mood

Fine-tune Qwen/Qwen3-8B on 65 bg pairs so background briefs match painter mood taste.

RL · Preference Pairs
narrative structure-19

Act Structure Gut

Fine-tune Qwen/Qwen3.5-9B on 75 act pairs so act breakdowns match showrunner structure taste.

RL · Preference Pairs
texture art-20

Texture Wear Sense

Fine-tune Qwen/Qwen3.5-4B on 55 texture pairs so wear notes match texture artist grime taste.

RL · Preference Pairs
cinematography-21

Camera Move Instinct

Fine-tune Qwen/Qwen3-8B on 60 camera pairs so move descriptions match DP instinct taste.

RL · Preference Pairs
comic adaptation-22

Comic Panel Flow

Fine-tune Qwen/Qwen3.5-4B on 50 panel pairs so panel breakdowns match comic adapter flow taste.

RL · Preference Pairs
render pipeline-23

Render Budget Taste

Fine-tune Qwen/Qwen3.5-9B on 70 render pairs so budget notes match pipeline lead tradeoff taste.

RL · Preference Pairs
character acting-24

Acting Pose Flair

Fine-tune Qwen/Qwen3-8B on 65 pose pairs so pose notes match animation director flair taste.

RL · Preference Pairs