{"record": {"aliases": ["clouda-ocr-benchmark"], "artifacts": [{"id": "gh:sahrasayed3-crypto/clouda-ocr-benchmark", "kind": "repo", "url": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark"}, {"kind": "website", "url": "https://cloudaocr.xyz"}], "categories": ["Language & Knowledge", "Software & AI Compute", "cs.AI"], "description": {"en": "Evaluates distorted Arabic document OCR on 177 page images with reference transcriptions using normalized Arabic character error rate and a primary leaderboard for models with full coverage."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark", "source": "claire_radar", "source_url": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark", "title": "Clouda OCR Benchmark"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:github:sahrasayed3-crypto/clouda-ocr-benchmark", "modality": null, "name": "Clouda OCR Benchmark", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-candidate", "display_eligible": "true", "origin_record_id": "bm_clouda-ocr-benchmark_4bff2fb7", "origin_source": "github", "recognition_confidence": "0.4", "record_sha256": "cfdbd870704f0de8bf06209284ce70dc5c41f8c81d1adb25e28a4d6a5210159a", "relation": "unclear", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-09-09T06:18:57.314550Z", "source_url": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark"}, "publisher": null, "released": "2026-09-07", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:github:sahrasayed3-crypto/clouda-ocr-benchmark", "source_url": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-github-sahrasayed3-crypto-clouda-ocr-benchmark", "source": "claire_radar", "source_benchmark_id": "github:sahrasayed3-crypto/clouda-ocr-benchmark", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["clouda-ocr-benchmark"], "annotation": "Mixed", "applicationDomains": ["Software & AI Compute"], "area": "Language & Knowledge", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark", "githubScope": "benchmark_repo", "githubStars": 1, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": null, "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"state": "not_applicable"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Specialized Arabic OCR benchmark with a clear leaderboard and public artifacts may attract moderate attention from NLP and OCR researchers in the first week.", "score": 45}, "capabilities": [], "confidence": "Medium", "construction": "Aggregate Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-09T06:18:57.314550Z", "inputHash": "126e3e94f4b63b966fa41ba0e3d6750c61193ab6cc22a67b00d1cd2594e0dc6f", "model": "deepseek-v4-pro"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "# Clouda OCR Benchmark — Public Release Arabic-first OCR and Document AI benchmark infrastructure from Clouda OCR.", "canonicalNameSource": "official_readme", "decisionReason": "The repository explicitly defines a 177-sample distorted Arabic OCR task, normalized CER scoring, and publishes stored outputs and evaluator code for external verification and reuse.", "evaluationEvidence": {"reuse": "All values above are recomputed offline from the stored per-sample outputs using the canonical evaluator; no OCR inference was rerun during reconciliation.", "scoring": "Primary metric **Normalized Arabic CER** (lower is better): character error rate computed after a deterministic Arabic normalization (NFC, alef/ya unification, tatweel and diacritic removal, whitespace collapse), averaged over each model's valid evaluated pairs.", "task": "Models transcribe **distorted Arabic document page images** (single-image, page-level OCR). Each of the 177 benchmark samples is a distorted page image plus a reference transcription."}, "model": "deepseek-v4-pro", "reviewedAt": "2026-09-09T06:18:57.314550Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-candidate", "demo": false, "description": "Evaluates distorted Arabic document OCR on 177 page images with reference transcriptions using normalized Arabic character error rate and a primary leaderboard for models with full coverage.", "discoverySources": [{"id": "github:sahrasayed3-crypto/clouda-ocr-benchmark", "type": "github", "url": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark"}], "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["discovered via github", "benchmark term in abstract", "evaluation protocol evidence", "public artifact URL", "no explicit benchmark release evidence"], "snippet": "clouda-ocr-benchmark Reproducible Arabic OCR benchmark by Clouda OCR — distorted documents, normalized Arabic CER, model comparisons, methodology, and public audit artifacts."}, "familyId": "bmf_3b96ad417b0d", "firstSeenAt": "2026-09-08", "heat": null, "id": "bm_clouda-ocr-benchmark_4bff2fb7", "indexedAt": "2026-09-08T06:12:01.862698Z", "industrySectors": ["Software & Cloud"], "links": {"code": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark", "project": "https://cloudaocr.xyz"}, "metrics": [], "motivation": "clouda-ocr-benchmark Reproducible Arabic OCR benchmark by Clouda OCR — distorted documents, normalized Arabic CER, model comparisons, methodology, and public audit artifacts.", "name": "Clouda OCR Benchmark", "oneLine": "Evaluates distorted Arabic document OCR on 177 page images with reference transcriptions using normalized Arabic character error rate and a primary leaderboard for models with full coverage.", "paperTitle": "clouda-ocr-benchmark", "primaryDomain": "Software & AI Compute", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-09-08T06:12:01.862698Z"}, "publications": [], "publishers": [{"name": "Clouda OCR", "organizationType": "company-research-lab", "role": "benchmark-publisher", "sourceUrl": "https://cloudaocr.xyz"}], "ranking": {"30d": {"confidence": "Low", "coverage": 0.55, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": 30, "historyDate": "2026-09-05", "rank": null, "score": null, "windowDays": 30}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.10220105459060908, "value": 1}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 111, "score": 28}, "method": "attention-ranking-v13", "rank": 111, "score": 28}, "90d": {"confidence": "Low", "coverage": 0.55, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": 0.09252268850709358, "value": 1}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 318, "score": 28}, "method": "attention-ranking-v13", "rank": 318, "score": 28}}, "readiness": "Runnable", "recognitionConfidence": 0.4, "relation": "unclear", "releasedAt": "2026-09-07", "source": {"authors": ["sahrasayed3-crypto"], "categories": [], "id": "github:sahrasayed3-crypto/clouda-ocr-benchmark", "publicSignals": {"githubStars": 0}, "title": "clouda-ocr-benchmark", "type": "github", "url": "https://github.com/sahrasayed3-crypto/clouda-ocr-benchmark"}, "sourceUpdatedAt": "2026-09-07T16:03:03Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "Provides a reproducible Arabic-first OCR benchmark with public outputs and metric code, addressing the need for transparent evaluation of Arabic document AI under realistic distortions."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
