Name: NLP_NER_Project
Author: manavshah409

Imported from manavshah409/NLP_NER_Project (AGENTS.md). Install upstream with npx skills add manavshah409/NLP_NER_Project. Copyright stays with the author.

IndicNewsNER research boundaries

The 100k classical CRF is frozen for Milestone 2C. Do not change its configuration, sample, binary, frozen source or environment without a separately authorised new version. Verify reports/crf/100k_baseline_freeze_manifest.json before final use.

During later model development or selection, do not read test predictions or illustrative examples in reports/crf/100k_final/*_predictions.jsonl or reports/crf/100k_final/restricted_examples/. Those files are for final reporting only. Use train_clean and validation_clean for model development. Do not tune any model from the final test metrics. New final test evaluations require explicit user authorisation. Never rerun a final evaluator to improve a score.

No 250k/full-data CRF, BiLSTM-CRF, IndicBERT, or English work is authorised as part of Milestone 2C. Preserve raw and frozen derived data.

Milestone 2D closure: future development must use the process guard described in docs/progress/MILESTONE_2D_CLOSURE.md. Install it before any data access. Direct historical training/evaluation commands are archival only; do not use them to bypass the guard. Only train_clean and validation_clean may be mounted in a future isolated development environment. Keep final aggregate metrics available for reporting but outside model-selection processes.

# after Install: the listing is in your workspace's routing pool - a plan picks it for its slot curl -s -X POST https://api.opensmartroute.ai/api/v1/route -H 'Authorization: Bearer $OSR_API_KEY' -H 'Content-Type: application/json' -d '{"text": "...", "plan": true}'

{ "ocm": "1", "id": "manavshah409-nlp-ner-project-nlp-ner-project", "kind": "skill", "name": "NLP_NER_Project", "description": "The 100k classical CRF is frozen for Milestone 2C. Do not change its configuration, sample, binary, frozen source or environment without a separately authorised new version. Verify `reports/crf/100k_baseline_freeze_manifest.json` before final use.", "publisher": "manavshah409", "version": "1.0.0", "capabilities": { "domains": [ "general" ], "tags": [ "skill-md", "github-agents" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "The 100k classical CRF is frozen for Milestone 2C. Do not change its configuration, sample, binary, frozen source or environment without a separately authorised new version. Verify `reports/crf/100k_baseline_freeze_manifest.json` before final use." ], "primary": false, "metadata": { "source": { "provider": "github-agents", "repository": "https://github.com/manavshah409/NLP_NER_Project", "path": "AGENTS.md", "ref": "1ccb7f998797adffdd79f69bf7e2d992d3dc7948", "url": "https://github.com/manavshah409/NLP_NER_Project/blob/1ccb7f998797adffdd79f69bf7e2d992d3dc7948/AGENTS.md", "key": "manavshah409/NLP_NER_Project/AGENTS.md" } }, "instructions": "# IndicNewsNER research boundaries\n\nThe 100k classical CRF is frozen for Milestone 2C. Do not change its configuration,\nsample, binary, frozen source or environment without a separately authorised new\nversion. Verify `reports/crf/100k_baseline_freeze_manifest.json` before final use.\n\nDuring later model development or selection, do not read test predictions or\nillustrative examples in `reports/crf/100k_final/*_predictions.jsonl` or\n`reports/crf/100k_final/restricted_examples/`. Those files are for final reporting\nonly. Use train_clean and validation_clean for model development. Do not tune any\n", "cost": { "context_tokens": 335 } }

NLP_NER_Project

About

IndicNewsNER research boundaries

Use it

Manifest

Reviews