JOB_OUTFITS.md

Professional outfits for existing adult identities

Snapshot: 16/4,640 sprites uploaded, 16 generated. Planned:464 existing adult human identities,19–95 years old,each with ten unique professional outfits. These images reuse personality IDs and do not add new identities. The entire character dataset target becomes 12,140 images across840 identities.

Every adult receives one job from each of ten families. Each of the100 professions has46 or47 images (60 professions have46;40 have47). Exactly equal counts are arithmetically impossible for464×10 images divided among100 jobs. Within every age/gender cell,each job's count also differs by at most1. Job allocation is independent of intended heritage and does not restrict occupations by gender. Age/heritage/gender describe intended prompt labels,not inferred appearance or real employment. All teen/child and special-creature categories are excluded.

All100 jobs

ID	Beruf	English title	Family	Planned sprites
job-001	Arzt/Ärztin	physician	health	47
job-002	Chirurg/Chirurgin	surgeon	health	47
job-003	Pflegefachkraft	nurse	health	46
job-004	Rettungssanitäter/in	paramedic	health	46
job-005	Apotheker/in	pharmacist	health	46
job-006	Zahnarzt/Zahnärztin	dentist	health	46
job-007	Medizinische Laborfachkraft	medical laboratory technician	health	47
job-008	Radiologie-Fachkraft	radiographer	health	46
job-009	Physiotherapeut/in	physiotherapist	health	47
job-010	Psychologe/Psychologin	psychologist	health	46
job-011	Feuerwehrkraft	firefighter	emergency	47
job-012	Polizeikraft	police officer	emergency	47
job-013	Kriminalermittler/in	detective	emergency	46
job-014	Sicherheitskraft	security guard	emergency	46
job-015	Küstenwache	coast guard officer	emergency	47
job-016	Rettungsschwimmer/in	lifeguard	emergency	47
job-017	Bergungs- und Suchrettungskraft	search and rescue worker	emergency	46
job-018	Katastrophenschutzkraft	civil defense worker	emergency	46
job-019	Park-Ranger/in	park ranger	emergency	46
job-020	Justizvollzugskraft	corrections officer	emergency	46
job-021	Kassierer/in	cashier	retail_hospitality	47
job-022	Supermarktmitarbeiter/in	supermarket worker	retail_hospitality	46
job-023	Bäcker/in	baker	retail_hospitality	46
job-024	Fleischer/in	butcher	retail_hospitality	46
job-025	Fischhändler/in	fishmonger	retail_hospitality	46
job-026	Florist/in	florist	retail_hospitality	47
job-027	Servicekraft im Restaurant	restaurant server	retail_hospitality	46
job-028	Barmann/Barfrau	bartender	retail_hospitality	47
job-029	Barista	barista	retail_hospitality	46
job-030	Hotel-Rezeptionist/in	hotel receptionist	retail_hospitality	47
job-031	Elektriker/in	electrician	construction	46
job-032	Installateur/in	plumber	construction	46
job-033	Zimmerer/Zimmerin	carpenter	construction	47
job-034	Maurer/in	bricklayer	construction	46
job-035	Maler/in	painter	construction	46
job-036	Dachdecker/in	roofer	construction	47
job-037	Fliesenleger/in	tiler	construction	47
job-038	Glaser/in	glazier	construction	47
job-039	Schlosser/in	locksmith	construction	46
job-040	Bauleiter/in	construction foreman	construction	46
job-041	Zerspanungsmechaniker/in	machinist	industrial_logistics	46
job-042	Schweißer/in	welder	industrial_logistics	46
job-043	Kfz-Mechaniker/in	auto mechanic	industrial_logistics	47
job-044	Klima- und Heizungstechniker/in	HVAC technician	industrial_logistics	46
job-045	Lagerfachkraft	warehouse worker	industrial_logistics	47
job-046	Gabelstaplerfahrer/in	forklift operator	industrial_logistics	46
job-047	Lkw-Fahrer/in	truck driver	industrial_logistics	47
job-048	Paketzusteller/in	delivery courier	industrial_logistics	47
job-049	Abfallwirtschaftskraft	sanitation worker	industrial_logistics	46
job-050	Bahn-Instandhaltungskraft	railway maintenance mechanic	industrial_logistics	46
job-051	Postzusteller/in	postal worker	civic_transport	46
job-052	Busfahrer/in	bus driver	civic_transport	46
job-053	Lokführer/in	train driver	civic_transport	46
job-054	Taxifahrer/in	taxi driver	civic_transport	46
job-055	Flugbegleiter/in	flight attendant	civic_transport	47
job-056	Linienpilot/in	airline pilot	civic_transport	47
job-057	Kommunale Instandhaltungskraft	municipal maintenance worker	civic_transport	47
job-058	Straßenreinigungskraft	street cleaner	civic_transport	47
job-059	Verkehrsüberwachungskraft	traffic warden	civic_transport	46
job-060	Verwaltungsangestellte/r	municipal clerk	civic_transport	46
job-061	Bankangestellte/r	bank teller	office_finance	46
job-062	Buchhalter/in	accountant	office_finance	47
job-063	Geldeintreiber/in	debt collection agent	office_finance	47
job-064	Versicherungsberater/in	insurance agent	office_finance	47
job-065	Immobilienmakler/in	real estate agent	office_finance	46
job-066	Rechtsanwalt/Rechtsanwältin	lawyer	office_finance	47
job-067	Richter/in	judge	office_finance	46
job-068	Finanzberater/in	financial adviser	office_finance	46
job-069	Büroleiter/in	office manager	office_finance	46
job-070	Übersetzer/in	translator	office_finance	46
job-071	Landwirt/in	farmer	outdoors_animals	47
job-072	Gärtner/in	gardener	outdoors_animals	46
job-073	Forstwirt/in	forester	outdoors_animals	47
job-074	Fischereifachkraft	fishery worker	outdoors_animals	46
job-075	Tierpfleger/in	animal caretaker	outdoors_animals	46
job-076	Tierarzt/Tierärztin	veterinarian	outdoors_animals	46
job-077	Landschaftsgärtner/in	landscape contractor	outdoors_animals	47
job-078	Winzer/in	viticulturist	outdoors_animals	46
job-079	Imker/in	beekeeper	outdoors_animals	46
job-080	Meeresbiologe/Meeresbiologin	marine biologist	outdoors_animals	47
job-081	Lehrer/in	school teacher	science_education	46
job-082	Hochschuldozent/in	university lecturer	science_education	47
job-083	Forschungswissenschaftler/in	research scientist	science_education	47
job-084	Astronom/in	astronomer	science_education	46
job-085	Archäologe/Archäologin	archaeologist	science_education	46
job-086	Vermessungstechniker/in	surveyor	science_education	46
job-087	Architekt/in	architect	science_education	46
job-088	Softwareentwickler/in	software developer	science_education	47
job-089	Journalist/in	journalist	science_education	47
job-090	Bibliothekar/in	librarian	science_education	46
job-091	Koch/Köchin	chef	creative_culture	47
job-092	Friseur/in	hairdresser	creative_culture	47
job-093	Barbier/in	barber	creative_culture	47
job-094	Schneider/in	tailor	creative_culture	46
job-095	Schuhmacher/in	shoe repairer	creative_culture	46
job-096	Fotograf/in	photographer	creative_culture	46
job-097	Tontechniker/in	sound technician	creative_culture	46
job-098	Bühnentechniker/in	stage technician	creative_culture	46
job-099	Museumskurator/in	museum curator	creative_culture	47
job-100	Musiklehrer/in	music teacher	creative_culture	46

The job CSV,full outfit allocation CSV,allocation checks and exact prompt/seed plan document every assignment. Clothing can suggest a profession but does not visually prove it; several civilian office roles naturally share similar clothes. Uniforms are fictional and unbranded,without weapons or readable insignia. Debt collection is represented by ordinary fully clothed civilian clothing and a calm expression,not violent conduct. Hair and face stay visible; protective headgear is intentionally omitted in these portrait assets.

Identity and inference

Each outfit edit independently references the same person's original neutral default PNG. The full-size reference checksum and source TAR/member are recorded; actual conditioning uses416×624 PIL BICUBIC. Only clothes change,with face,hair,skin tone,body proportions,pose,framing,lighting and a neutral slightly friendly expression explicitly preserved. Small model-dependent changes can remain. All garments are opaque and fully covering; greens and teals are avoided for chroma keying. Outputs use the selected warm-painted Ghibli-inspired style,832×1,248 RGB,50 real-CFG4 steps,BF16,the undistilled FLUX.2-klein-base-9B and the linked rank64 character LoRA.

New TARs live in job-data/jobs-gN-NNNNN.tar,typically20 images with smaller pilot/final shards. Each stable sample ID <personality_id>-job-NNN has adjacent .png,.json and .txt members. JSON includes profession,base/planned outfit,identity,prompt/negative prompt,seed,reference SHA256/source location,actual inference parameters and timing. jobs isolates this family after the first upload; default includes all uploaded families. Identity-based splits must keep every emotion/outfit of the same personality ID together.

from datasets import load_dataset
ds=load_dataset("laion/Vivarium-Greenscreen-Characters-FLUX-9B","jobs",split="train",streaming=True)
sample=next(iter(ds))
print(sample["json"]["personality_id"],sample["json"]["job_title_en"])

Portable inference keeps reference bytes in memory and writes a TAR:

python tools/infer_job_outfit_tar.py --id person-0281-job-001 --output physician.tar

Use the original pinned dependencies and model setup. The generator can load the neutral reference from the public source TAR using the checksum/range index or accept --reference with its PNG. Progress,shard manifests,inference records,review and integrity reports are published under metadata/job-outfits/. The first16 shared-GPU/CPU-offload pilot timings are separated from dedicated-worker throughput.