IssueAggregator

[25 MRG] Sign pack: `sister` (family) — photo/video + landmarks

[25 MRG] Sign pack: `brother` (family) — photo/video + landmarks

[25 MRG] Sign pack: `father` (family) — photo/video + landmarks

[25 MRG] Sign pack: `mother` (family) — photo/video + landmarks

[25 MRG] Sign pack: `which` (question) — photo/video + landmarks

[25 MRG] Sign pack: `how` (question) — photo/video + landmarks

[25 MRG] Sign pack: `why` (question) — photo/video + landmarks

[25 MRG] Sign pack: `who` (question) — photo/video + landmarks

[25 MRG] Sign pack: `when` (question) — photo/video + landmarks

[25 MRG] Sign pack: `where` (question) — photo/video + landmarks

[25 MRG] Sign pack: `what` (question) — photo/video + landmarks

[25 MRG] Sign pack: `they` (pronoun) — photo/video + landmarks

[25 MRG] Sign pack: `we` (pronoun) — photo/video + landmarks

[25 MRG] Sign pack: `you` (pronoun) — photo/video + landmarks

[25 MRG] Sign pack: `me_i` (pronoun) — photo/video + landmarks

[25 MRG] Sign pack: `name` (identity) — photo/video + landmarks

[25 MRG] Sign pack: `bad` (core) — photo/video + landmarks

[25 MRG] Sign pack: `good` (core) — photo/video + landmarks

[25 MRG] Sign pack: `fine_ok` (core) — photo/video + landmarks

[25 MRG] Sign pack: `tired` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `scared` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `angry` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `sad` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `happy` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `love` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `like` (emotion) — photo/video + landmarks

[25 MRG] Sign pack: `need` (core) — photo/video + landmarks

[25 MRG] Sign pack: `want` (core) — photo/video + landmarks

[25 MRG] Sign pack: `goodbye` (greetings) — photo/video + landmarks

[50 MRG] Webcam demo: OpenCV capture loop → live gloss print

[50 MRG] Gloss→sentence: template pack + simple NLG for multi-gloss

[50 MRG] FastAPI: POST /infer/text and /infer/voice

[50 MRG] Sign→voice: pyttsx3 or edge-tts real backend

[50 MRG] Sign→text: sequence model baseline (CTC or tiny transformer stub)

[50 MRG] MediaPipe Holistic landmark extractor adapter

[25 MRG] Sequence schema validation with pydantic

[25 MRG] Gloss vocab: load from YAML/JSON + expand DEFAULT_GLOSS

[25 MRG] CLI: loru eval toy — accuracy table over data/samples

[25 MRG] Docs: DATASETS.md catalog of public sign corpora + licenses