File size: 8,719 Bytes
f9bc0c1 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 | {
"version": "eval-target-coverage-v1",
"updated": "2026-07-15",
"scope_note": "Tracks intended Nepal/diaspora OCR evaluation coverage. Claim readiness is determined by linked eval-pack normalization recipes, eval run recipes, and their blockers; target presence alone is not evidence.",
"target_sets": [
{
"id": "limbu-first-two-script-ocr",
"description": "Required coverage for the Limbu-first OCR pipeline: Sirijonga-script Limbu and Devanagari-written Limbu must both be represented by claim-ready eval harness entries before any two-script Limbu OCR claim.",
"required_for_claim": true,
"requires_page_eval_recipe": true,
"requires_book_candidate_evidence": true,
"targets": [
{
"language": "limbu",
"display_name": "Limbu/Yakthung",
"script": "Limbu",
"first_script_target": "Limbu/Sirijonga",
"eval_status": "registered_real_pack",
"normalization_pack_id": "limbu-gorkhapatra-real-v1",
"eval_recipe_id": "limbu-sirijonga-gorkhapatra-oracle-crop",
"page_eval_recipe_id": "limbu-sirijonga-gorkhapatra-book-conf070-coverage-gated",
"notes": "Tracks the current real Sirijonga recognizer pack and the operationally admitted one-page book replay. Still blocked by machine-converted GT, row claim eligibility, one-page sample size, unseen-selection status, and missing packaged model-card evidence."
},
{
"language": "limbu",
"display_name": "Limbu/Yakthung",
"script": "Devanagari",
"first_script_target": "Devanagari-written Limbu",
"eval_status": "registered_real_pack",
"normalization_pack_id": "limbu-devanagari-cdc-line-crop-corpus-draft-v1",
"eval_recipe_id": "limbu-devanagari-cdc-ppocrv5-devclean-heldout",
"page_eval_recipe_id": "limbu-devanagari-cdc-page-marker-v2-champion-coverage-gated",
"notes": "Tracks the current Devanagari-written Limbu CDC book-page candidate: page-marker-v2 with champion non-regression and unresolved-line coverage admission. It is still not claim-ready until the pack is frozen and text/line references are reviewed."
}
]
},
{
"id": "first-ten-document-ocr",
"description": "First ten language targets from docs/NEPAL_MULTILINGUAL_OCR_GOAL.md.",
"required_for_claim": false,
"targets": [
{
"language": "limbu",
"display_name": "Limbu/Yakthung",
"script": "Limbu",
"first_script_target": "Limbu/Sirijonga",
"eval_status": "partial_real_recognizer_pack",
"normalization_pack_id": "limbu-gorkhapatra-real-v1",
"eval_recipe_id": "limbu-sirijonga-gorkhapatra-oracle-crop",
"notes": "First non-Devanagari target; current pack lacks row provenance and native-reader calibration."
},
{
"language": "nepali",
"display_name": "Nepali",
"script": "Devanagari",
"first_script_target": "Devanagari",
"eval_status": "missing_real_eval_pack",
"notes": "The local 79,364-label shared harvest is a training inventory with 39,006 missing images, not a held-out Nepali eval pack."
},
{
"language": "english",
"display_name": "English",
"script": "Latin",
"first_script_target": "Latin",
"eval_status": "missing_real_eval_pack",
"notes": "Must not regress for mixed documents, but no claim-grade recognizer eval pack is registered here."
},
{
"language": "tamang",
"display_name": "Tamang",
"script": "Devanagari",
"first_script_target": "Devanagari first; Tamyig audit later",
"eval_status": "missing_real_eval_pack",
"notes": "Local corpus exists per project docs; no registered real OCR eval pack yet."
},
{
"language": "magar",
"display_name": "Magar",
"script": "Devanagari",
"first_script_target": "Devanagari first",
"eval_status": "missing_real_eval_pack",
"notes": "Gorkhapatra item confirmed in project docs; no registered real OCR eval pack yet."
},
{
"language": "wambule",
"display_name": "Wambule",
"script": "Devanagari",
"first_script_target": "Devanagari first",
"eval_status": "missing_real_eval_pack",
"notes": "Gorkhapatra item confirmed in project docs; no registered real OCR eval pack yet."
},
{
"language": "bantawa_rai",
"display_name": "Bantawa/Rai",
"script": "Devanagari",
"first_script_target": "Devanagari or Kirat Rai depending source",
"eval_status": "missing_real_eval_pack",
"notes": "Rai-family target; Kirat Rai native-script pack is tracked separately as current real-pack coverage."
},
{
"language": "kulung",
"display_name": "Kulung",
"script": "Devanagari",
"first_script_target": "Devanagari first",
"eval_status": "missing_real_eval_pack",
"notes": "Gorkhapatra item confirmed in project docs; no registered real OCR eval pack yet."
},
{
"language": "yakkha",
"display_name": "Yakkha",
"script": "Devanagari",
"first_script_target": "Devanagari first",
"eval_status": "missing_real_eval_pack",
"notes": "Gorkhapatra item confirmed in project docs; no registered real OCR eval pack yet."
},
{
"language": "tharu_rana",
"display_name": "Tharu/Rana Tharu",
"script": "Devanagari",
"first_script_target": "Devanagari first",
"eval_status": "missing_real_eval_pack",
"notes": "Multiple varieties; labels must remain explicit before any claim."
}
]
},
{
"id": "current-real-recognizer-packs",
"description": "Current real recognizer eval packs registered for the harness.",
"required_for_claim": true,
"targets": [
{
"language": "newa",
"display_name": "Newa / Nepal Bhasa",
"script": "Newa",
"first_script_target": "Prachalit/Newa",
"eval_status": "registered_real_pack",
"normalization_pack_id": "newa-zenodo-6967421-v1",
"eval_recipe_id": "newa-zenodo-native-retrain-oracle-crop"
},
{
"language": "sunuwar",
"display_name": "Sunuwar / Jenticha",
"script": "Sunuwar",
"first_script_target": "Jenticha/Sunuwar",
"eval_status": "missing_real_eval_pack",
"notes": "The historical 429-row pack and checkpoint are absent, and source redistribution permission is unresolved."
},
{
"language": "lepcha",
"display_name": "Lepcha",
"script": "Lepcha",
"first_script_target": "Lepcha",
"eval_status": "partial_real_recognizer_pack",
"normalization_pack_id": "lepcha-herald-real-v1",
"eval_recipe_id": "lepcha-herald-real-oracle-crop",
"notes": "132 rows / 107 unique labels exist, but the pack is unfrozen and lacks canonical provenance, edition-separation evidence, native review, and a checkpoint."
},
{
"language": "limbu",
"display_name": "Limbu/Yakthung",
"script": "Limbu",
"first_script_target": "Limbu/Sirijonga",
"eval_status": "registered_real_pack",
"normalization_pack_id": "limbu-gorkhapatra-real-v1",
"eval_recipe_id": "limbu-sirijonga-gorkhapatra-oracle-crop"
},
{
"language": "kiratrai",
"display_name": "Kirat Rai",
"script": "KiratRai",
"first_script_target": "Kirat Rai",
"eval_status": "partial_real_recognizer_pack",
"normalization_pack_id": "kiratrai-herald-real-v1",
"eval_recipe_id": "kiratrai-herald-real-oracle-crop",
"notes": "All 53 unique labels in the 70-row pack occur in the real-render training set; it is not held out and no checkpoint is present."
},
{
"language": "devanagari_shared",
"display_name": "Shared Devanagari Gorkhapatra",
"script": "Devanagari",
"first_script_target": "Shared Devanagari",
"eval_status": "missing_real_eval_pack",
"notes": "The local shared harvest is a training inventory with 79,364 labels and only 40,358 images, not a frozen held-out eval pack."
}
]
}
]
}
|