{ "registry_id": "nepal-v1", "description": "Nepal multilingual OCR target registry. Limbu/Sirijonga is the first non-Devanagari target; English is included for preservation but does not count toward the Nepal-language target.", "minimum_nepal_language_count": 10, "created_for": "PaddleOCR script-aware fine-tuning, synthesis, and real-document validation", "claim_policy": "Synthetic data is allowed for training and ablations only. SOTA claims require held-out real/public validation artifacts.", "languages": [ { "id": "limbu", "name": "Limbu / Yakthung", "endonym": "ᤕᤠᤰᤌᤢᤱ", "iso_639_3": "lif", "priority": 1, "counts_toward_nepal_language_target": true, "first_target": true, "scripts": [ { "script": "Limbu / Sirijonga", "primary": true, "status": "encoded", "unicode_ranges": [ "U+1900..U+194F" ], "notes": "First non-Devanagari OCR target. Requires renderability and dictionary audits before training admission." }, { "script": "Devanagari", "primary": false, "status": "fallback_transcription", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Useful for parallel/transcribed resources, not the primary Limbu-script claim target." } ], "font_candidates": [ { "name": "Namdhinggo Regular", "path": "data/fonts/namdhinggo/Namdhinggo-3.100/Namdhinggo-Regular.ttf", "status": "primary_sirijonga_font_render_audit_passed" }, { "name": "Noto Serif Devanagari", "path": "data/fonts/google-fonts-2026-06-16/notoserifdevanagari/NotoSerifDevanagari-VariableFont_wdth,wght.ttf", "status": "secondary_devanagari_font_render_audit_passed" }, { "name": "Noto Sans Limbu", "path": "/System/Library/Fonts/Supplemental/NotoSansLimbu-Regular.ttf", "status": "pending_render_audit" }, { "name": "local-ltk-fonts", "path": "/Users/cdjk/github/llm/ltk/fonts.zip", "status": "legacy_archive_pending_coverage_audit" } ], "paddle_dictionary": "data/dictionaries/limbu-sirijonga-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_limbu_letters_and_signs" }, "corpora": [ { "name": "LTK Limbu resources", "path": "/Users/cdjk/github/llm/ltk/limbu", "host": "cdjk", "type": "local_verified_directory", "status": "schema_audit_pending" }, { "name": "Limdic dictionary", "path": "/Users/cdjk/github/llm/ltk/limbu/data/limdic_dictionary.json", "host": "cdjk", "type": "lexicon", "status": "exists_local" }, { "name": "Pangloss dictionary TSV", "path": "/Users/cdjk/github/llm/ltk/limbu/data/pangloss_dictionary.tsv", "host": "cdjk", "type": "lexicon", "status": "exists_local" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal article images", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "pending_manual_label" }, { "name": "Gorkhapatra epaper PDFs", "path": "data/gorkhapatra-pack-smoke", "type": "real_document_validation_pack", "status": "pending_manual_label" } ], "gorkhapatra_language_labels": [ "लिम्बु" ], "notes": "Primary research target for low-resource script-aware PaddleOCR adaptation." }, { "id": "nepali", "name": "Nepali", "endonym": "नेपाली", "iso_639_3": "nep", "priority": 2, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Established base script and regression-protection target." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/nepali-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Nepali hunspell dictionary", "path": "/Users/cdjk/github/llm/ltk/ne_NP.dic", "host": "cdjk", "type": "lexicon", "status": "exists_local" } ], "validation_sources": [ { "name": "Gorkhapatra epaper PDFs", "path": "data/gorkhapatra-pack-smoke", "type": "real_document_validation_pack", "status": "pending_manual_label" } ], "gorkhapatra_language_labels": [ "नेपाली" ], "notes": "Must not regress versus stock PaddleOCR/Tesseract/Surya baselines." }, { "id": "tamang", "name": "Tamang", "endonym": "तामाङ", "iso_639_3": "taj", "priority": 3, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari resources; dialect coverage must be tracked separately." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/tamang-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "LTK Tamang resources", "path": "/Users/cdjk/github/llm/ltk/tamang", "host": "cdjk", "type": "local_verified_directory", "status": "schema_audit_pending" }, { "name": "Western Tamang dictionary TSV", "path": "/Users/cdjk/github/llm/ltk/tamang/data/western_tamang_dictionary.tsv", "host": "cdjk", "type": "lexicon", "status": "exists_local" }, { "name": "NepTam80K train/test CSV", "path": "/Users/cdjk/github/llm/ltk/tamang/data", "host": "cdjk", "type": "parallel_or_synthetic_text", "status": "schema_audit_pending" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_audit_pending" } ], "gorkhapatra_language_labels": [ "तामाङ" ], "notes": "High-priority Devanagari low-resource language with existing local resources." }, { "id": "magar", "name": "Magar", "endonym": "मगर", "iso_639_3": "mrd", "priority": 4, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/magar-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "LTK Magar resources", "path": "/Users/cdjk/github/llm/ltk/magar", "host": "cdjk", "type": "local_verified_directory", "status": "schema_audit_pending" }, { "name": "Western Magar dictionary TSV", "path": "/Users/cdjk/github/llm/ltk/magar/data/western_magar_dictionary.tsv", "host": "cdjk", "type": "lexicon", "status": "exists_local" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "मगर" ], "notes": "Seen in Naya Nepal audit labels; needs manual reference pack." }, { "id": "wambule", "name": "Wambule", "endonym": "वाम्बुले", "iso_639_3": "wme", "priority": 5, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/wambule-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Wambule articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "वाम्बुले" ], "notes": "Real-source first; no local training corpus audited yet." }, { "id": "rana-tharu", "name": "Rana Tharu", "endonym": "राना थारु", "iso_639_3": "thr", "priority": 6, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/rana-tharu-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Rana Tharu articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "राना थारु", "थारु" ], "notes": "Track Rana Tharu separately from broader Tharu labels when references are created." }, { "id": "bantawa", "name": "Bantawa / Rai", "endonym": "बान्तावा", "iso_639_3": "bap", "priority": 7, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." }, { "script": "Kirat Rai", "primary": false, "status": "requires_unicode_font_audit", "unicode_ranges": [ "U+16D40..U+16D7F" ], "notes": "Potential future script target; do not train or claim until font/runtime support is verified." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Kanchenjunga Regular", "path": "data/fonts/google-fonts-2026-06-16/kanchenjunga/Kanchenjunga-Regular.ttf", "status": "kirat_rai_font_render_audit_passed" }, { "name": "Kanchenjunga Bold", "path": "data/fonts/google-fonts-2026-06-16/kanchenjunga/Kanchenjunga-Bold.ttf", "status": "kirat_rai_font_render_audit_passed" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/bantawa-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Bantawa articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "बान्तावा" ], "notes": "Rai-language coverage must stay language-specific; do not collapse all Rai languages into one metric." }, { "id": "kulung", "name": "Kulung", "endonym": "कुलुङ", "iso_639_3": "kle", "priority": 8, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/kulung-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Kulung articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "कुलुङ" ], "notes": "Real-source first; needs corpus expansion." }, { "id": "yakkha", "name": "Yakkha", "endonym": "याक्खा", "iso_639_3": "ybh", "priority": 9, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/yakkha-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Yakkha articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "याक्खा" ], "notes": "Real-source first; needs corpus expansion." }, { "id": "bhojpuri", "name": "Bhojpuri", "endonym": "भोजपुरी", "iso_639_3": "bho", "priority": 10, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/bhojpuri-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Bhojpuri articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "भोजपुरी" ], "notes": "Large-language baseline for Devanagari multilingual transfer." }, { "id": "santhali", "name": "Santhali", "endonym": "सन्थाली", "iso_639_3": "sat", "priority": 11, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Ol Chiki", "primary": true, "status": "encoded", "unicode_ranges": [ "U+1C50..U+1C7F" ], "notes": "Own-script target after Limbu; needs font/render audit." }, { "script": "Devanagari", "primary": false, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "May appear in Devanagari-transcribed sources." } ], "font_candidates": [ { "name": "Noto Sans Ol Chiki", "path": "data/fonts/NotoSansOlChiki-Regular.ttf", "status": "pending_render_audit" }, { "name": "Noto Sans Ol Chiki Variable", "path": "data/fonts/google-fonts-2026-06-16/notosansolchiki/NotoSansOlChiki-VariableFont_wght.ttf", "status": "ol_chiki_font_render_audit_passed" }, { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/santhali-ol-chiki-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_ol_chiki_letters" }, "corpora": [ { "name": "Gorkhapatra/Naya Nepal Santhali articles", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_text_pending", "status": "needs_extraction" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "सन्थाली" ], "notes": "Second own-script candidate after Limbu if real rendered sources and fonts are available." }, { "id": "maithili", "name": "Maithili", "endonym": "मैथिली", "iso_639_3": "mai", "priority": 12, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." }, { "script": "Tirhuta", "primary": false, "status": "future_audit", "unicode_ranges": [ "U+11480..U+114DF" ], "notes": "Own-script support is a future target and not part of the v1 claim without real validation." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Noto Sans Tirhuta", "path": "data/fonts/google-fonts-2026-06-16/notosanstirhuta/NotoSansTirhuta-Regular.ttf", "status": "tirhuta_font_render_audit_passed" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/maithili-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Maithili corpus discovery", "path": "TBD", "type": "external_corpus_pending", "status": "not_audited" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal or public documents", "path": "TBD", "type": "real_document_discovery", "status": "not_audited" } ], "gorkhapatra_language_labels": [ "मैथिली" ], "notes": "Included because it is a major Nepal language; needs source discovery before training." }, { "id": "gurung", "name": "Gurung", "endonym": "गुरुङ", "iso_639_3": "gvr", "priority": 13, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari; language-specific source claims require separate validation." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/gurung-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Bible Brain Gurung text manifest", "path": "/Volumes/TRANSCEND/bible_brain_nepal/gvr/GVRWBT_Gurung/metadata/text_manifest.jsonl", "host": "cdjk@100.110.148.43", "type": "real_text_inventory", "status": "schema_verified_by_bible_brain_converter" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "गुरुङ" ], "notes": "Remote Bible Brain text is training/ablation input only until license review and held-out real documents are verified." }, { "id": "dhimal", "name": "Dhimal", "endonym": "धिमाल", "iso_639_3": "dhi", "priority": 14, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/dhimal-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Bible Brain Dhimal text manifest", "path": "/Volumes/TRANSCEND/bible_brain_nepal/dhi/DHIDPI_Dhimal/metadata/text_manifest.jsonl", "host": "cdjk@100.110.148.43", "type": "real_text_inventory", "status": "schema_verified_by_bible_brain_converter" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "धिमाल" ], "notes": "Remote Bible Brain text is training/ablation input only; needs Gorkhapatra page references for claims." }, { "id": "sunuwar", "name": "Sunuwar", "endonym": "सुनुवार", "iso_639_3": "suz", "priority": 15, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari; any community-script support is out of v1 scope until audited." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/sunuwar-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Bible Brain Sunuwar text manifest", "path": "/Volumes/TRANSCEND/bible_brain_nepal/suz/SUZWBT_Sunuwar/metadata/text_manifest.jsonl", "host": "cdjk@100.110.148.43", "type": "real_text_inventory", "status": "schema_verified_by_bible_brain_converter" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "सुनुवार" ], "notes": "Remote Bible Brain text is training/ablation input only; needs held-out real document references." }, { "id": "rajbanshi", "name": "Rajbanshi", "endonym": "राजवंशी", "iso_639_3": "rjs", "priority": 16, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Initial OCR route uses Devanagari." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/rajbanshi-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Bible Brain Rajbanshi text manifest", "path": "/Volumes/TRANSCEND/bible_brain_nepal/rjs/RJSMTS_Rajbanshi/metadata/text_manifest.jsonl", "host": "cdjk@100.110.148.43", "type": "real_text_inventory", "status": "schema_verified_by_bible_brain_converter" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "राजवंशी" ], "notes": "Remote Bible Brain text is training/ablation input only; real publication validation remains required." }, { "id": "sherpa", "name": "Sherpa", "endonym": "शेर्पा", "iso_639_3": "xsr", "priority": 17, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "notes": "Bible Brain and Gorkhapatra-visible text paths are Devanagari/transcription first; Tibetan-script support is separate future work." } ], "font_candidates": [ { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit" }, { "name": "Devanagari MT", "path": "/System/Library/Fonts/Supplemental/DevanagariMT.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/sherpa-devanagari-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes" }, "corpora": [ { "name": "Bible Brain Sherpa text manifest", "path": "/Volumes/TRANSCEND/bible_brain_nepal/xsr/XSRXSR_Sherpa/metadata/text_manifest.jsonl", "host": "cdjk@100.110.148.43", "type": "real_text_inventory", "status": "schema_verified_by_bible_brain_converter" } ], "validation_sources": [ { "name": "Gorkhapatra/Naya Nepal", "path": "outputs/gorkhapatra-source-audit", "type": "real_document_discovery", "status": "language_label_seen" } ], "gorkhapatra_language_labels": [ "शेर्पा" ], "notes": "Remote Bible Brain text is training/ablation input only; script status must be kept explicit." }, { "id": "english", "name": "English", "endonym": "English", "iso_639_3": "eng", "priority": 99, "counts_toward_nepal_language_target": false, "first_target": false, "scripts": [ { "script": "Latin", "primary": true, "status": "encoded", "unicode_ranges": [ "U+0000..U+007F", "U+0080..U+00FF" ], "notes": "Regression-protection target only." } ], "font_candidates": [ { "name": "Arial Unicode MS", "path": "/Library/Fonts/Arial Unicode.ttf", "status": "pending_render_audit" }, { "name": "Helvetica", "path": "/System/Library/Fonts/Helvetica.ttc", "status": "pending_render_audit" } ], "paddle_dictionary": "data/dictionaries/english-latin-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_ascii_and_common_latin" }, "corpora": [ { "name": "NVIDIA OCR Synthetic Multilingual English subset", "path": "huggingface:nvidia/OCR-Synthetic-Multilingual-v1", "type": "synthetic_line_crops", "status": "schema_audit_required_before_use" } ], "validation_sources": [ { "name": "Stock English OCR eval subset", "path": "TBD", "type": "regression_eval", "status": "not_audited" } ], "gorkhapatra_language_labels": [], "notes": "Included to protect English accuracy in mixed documents; not counted as a Nepal-language target." }, { "id": "newar", "name": "Newar / Nepal Bhasa", "endonym": "नेपाल भाषा", "iso_639_3": "new", "priority": 18, "counts_toward_nepal_language_target": true, "first_target": false, "scripts": [ { "script": "Devanagari", "profile_id": "devanagari", "aliases": [ "Deva", "Devanagari script", "देवनागरी" ], "primary": true, "status": "encoded", "source_unicode_status": "encoded", "unicode_script": "Devanagari", "unicode_ranges": [ "U+0900..U+097F", "U+A8E0..U+A8FF" ], "ocr_support_status": "operational_recovered_model_diagnostic_only", "input_policy": "unicode_text_or_source_image", "output_policy": "native_devanagari_unicode", "notes": "The recovered deva-v2 export is available for diagnostic Devanagari OCR with exact-hash custody. Historical Gorkhapatra/Newari labels and scores remain claim-ineligible because rights, row provenance, native review, leakage, and split-independence gates fail." }, { "script": "Newar / Prachalit", "profile_id": "newa_prachalit", "aliases": [ "Newa", "Newar script", "Newari script", "Nepal Lipi", "Nepaalalipi", "Prachalit" ], "primary": false, "status": "encoded", "source_unicode_status": "encoded", "unicode_script": "Newa", "unicode_ranges": [ "U+11400..U+1147F" ], "ocr_support_status": "operational_recovered_model_diagnostic_failed_fit_only", "input_policy": "unicode_text_or_source_image", "output_policy": "native_newa_unicode", "notes": "Own-script target (second focus language per RESEARCH.md 2026-07-04). The recovered native-v1 checkpoint has an exact CPU inference export and one network-denied CPU smoke on a synthetic Newa line, but it remains only a known failed-fit diagnostic: the original PPLCNetV3 path supplied 40 CTC timesteps while 95.7% of training labels were longer. The smoke performed no output-to-source comparison or scoring, and no real-domain codepoint CER is admitted; the isolated CTC200 retrain and native Cambridge review remain open." }, { "script": "Bhujimol / Bhujinmol", "profile_id": "bhujimol", "aliases": [ "Bhujimol", "Bhujinmol", "Bhujimol script", "Bhujinmol script" ], "primary": false, "status": "newa_style_profile_no_separate_script_encoding", "source_unicode_status": "no_separate_script_encoding", "unicode_script": "Newa", "unicode_ranges": [ "U+11400..U+1147F" ], "ocr_support_status": "profile_registered_no_model_font_or_admitted_dataset", "input_policy": "source_image_or_newa_text_with_explicit_bhujimol_style", "output_policy": "newa_unicode_transcription_with_source_style_metadata", "notes": "A separate historical visual/headstroke profile. Unicode expert document N4555 describes possible unification with Newa for plain text, but OCR still requires Bhujimol-specific pixels, fonts, review, and evaluation. A CC0 basic-glyph vector chart is available for reviewer reference; it is not a font or OCR dataset.", "reference_assets": [ { "name": "Commons Bhujimol Unicode basic-glyph vector chart", "kind": "cc0_basic_glyph_vector_chart_not_font", "source_page": "https://commons.wikimedia.org/wiki/File:Bhujimol_script_characters_(Unicode).svg", "source_file": "https://upload.wikimedia.org/wikipedia/commons/4/41/Bhujimol_script_characters_%28Unicode%29.svg", "sha256": "bf79c86f6156b6630f9480a04f1fe33ffce0965edbea62be9e9339ad10d8655d", "license": "CC0 1.0", "status": "audited_reference_only_no_font_or_training_admission", "explicit_machine_readable_codepoint_mapping": false, "shaping_rules_present": false, "training_payload_admitted": false } ] }, { "script": "Ranjana", "profile_id": "ranjana", "aliases": [ "Ranjana", "Ranjana script", "Ranjana Lipi", "Rañjana", "Rañjanā" ], "primary": false, "status": "unencoded", "source_unicode_status": "unencoded", "unicode_script": null, "unicode_ranges": [], "transcription_unicode_script": "Newa", "transcription_unicode_ranges": [ "U+11400..U+1147F" ], "ocr_support_status": "image_profile_registered_unencoded_no_model_or_admitted_dataset", "input_policy": "source_image_or_audited_legacy_font_mapping_only", "output_policy": "newa_unicode_transcription_with_source_image_and_ranjana_style_metadata", "notes": "Ranjana remains unencoded in Unicode 17.0; the 2023 Unicode document is a preliminary proposal. Support therefore means a distinct image profile and explicit transcription policy, never a fabricated Ranjana Unicode block." } ], "font_candidates": [ { "name": "Noto Sans Newa", "path": "data/fonts/all/newa/NotoSansNewa-Regular.ttf", "status": "newa_font_used_in_synthesis", "writing_system_profiles": [ "newa_prachalit" ] }, { "name": "Nepal Lipi Unicode", "path": "data/fonts/all/newa/NepalLipiUnicode.ttf", "status": "newa_font_used_in_synthesis", "writing_system_profiles": [ "newa_prachalit" ] }, { "name": "Nithya Ranjana DU", "path": "data/fonts/all/devanagari/NithyaRanjanaDU-Regular.otf", "status": "audited_ofl_devanagari_mapped_ranjana_reference_only", "sha256": "8a2ebe626740270ac62c52815f2237326cdc0ed137041e833b6d404f9771b92b", "license": "SIL Open Font License, Version 1.1", "source_release": "https://github.com/EkType/Nithya-Ranjana/releases/tag/1.000", "mapping_basis": "devanagari_unicode", "training_payload_admitted": false, "writing_system_profiles": [ "ranjana" ] }, { "name": "Nithya Ranjana NU", "path": "data/fonts/all/newa/NithyaRanjanaNU-Regular.otf", "status": "audited_ofl_newa_mapped_ranjana_render_candidate", "sha256": "22a12a91e119953d45eb07095376db4ffdd23904112c53ead1dfab9906691528", "license": "SIL Open Font License, Version 1.1", "source_release": "https://github.com/EkType/Nithya-Ranjana/releases/tag/1.000", "mapping_basis": "newa_unicode", "training_payload_admitted": false, "writing_system_profiles": [ "ranjana" ] }, { "name": "Noto Sans Devanagari", "path": "data/fonts/NotoSansDevanagari-VariableFont_wdth,wght.ttf", "status": "pending_render_audit", "writing_system_profiles": [ "devanagari" ] } ], "paddle_dictionary": "data/dictionaries/newa-native-v1.txt", "normalization": { "unicode_form": "NFC", "policy": "preserve_devanagari_graphemes_and_newa_letters" }, "corpora": [ { "name": "Gorkhapatra Newar silver line packs (Devanagari)", "path": "outputs/language-pack-silver-lines-2026-07-04 + outputs/language-pack-silver-lines-batch2-2026-07-04 (slug label-1fefbafa97)", "type": "real_document_silver_lines", "status": "harvested_2026_07_04_machine_converted" }, { "name": "Leipzig Newari wikipedia sentences", "path": "data/leipzig-newari", "type": "text_corpus", "status": "exists_local" }, { "name": "Zenodo Newa manuscript crops", "path": "data/newa-zenodo", "type": "real_handwritten_crops", "status": "exists_local_cross_domain" } ], "validation_sources": [ { "name": "Gorkhapatra Naya Nepal 'नेपाल' language pages", "path": "outputs/language-pack-inventory-merged-2026-07-04", "type": "real_document_discovery", "status": "harvested_pending_frozen_split" } ], "gorkhapatra_language_labels": [ "नेपाल" ], "notes": "Newari support is tracked as four explicit writing-system profiles: Newar/Prachalit, Bhujimol, Ranjana, and Devanagari. Profile registration is not a model-quality claim; each visual system needs separately admitted pixels, labels, leakage controls, and real-domain evaluation." } ] }