Spaces:

aditya0103
/

structured-data-extractor

Sleeping

App Files Files Community

structured-data-extractor / scripts

49.5 kB

Ctrl+K

Ctrl+K

2 contributors

History: 7 commits

aditya0103's picture

v4: fine-tuning pipeline complete + bug fixes

d95323c 4 days ago

build_filings_gt.py

8.53 kB
v2.2: filing prompt uplift + XBRL FIN_MAP expansion + diagnose-loop narrative 6 days ago
compare_finetune.py

6.85 kB
v4: fine-tuning pipeline complete + bug fixes 4 days ago
download_edgar.py

8.33 kB
v2.2: filing prompt uplift + XBRL FIN_MAP expansion + diagnose-loop narrative 6 days ago
launch_finetune.py

6.43 kB
v4: fine-tuning pipeline complete + bug fixes 4 days ago
prep_datasets.py

3.79 kB
v2.1: SEC 10-K live eval (0.56 F1 at 0.06/doc) + CI ruff fixes 7 days ago
prep_ft_dataset.py

6.84 kB
v4: fine-tuning pipeline complete + bug fixes 4 days ago
run_eval.py

517 Bytes
Tasks 5-7: eval harness, FastAPI backend, Paper & Ink UI- src/eval/ — precision/recall/F1 harness with type-aware comparators,micro/macro F1, CSV + markdown reports, --model benchmark flag- src/api/ — FastAPI backend with /extract, /schemas, /health,request-ID middleware, typed error envelope, injectable extractor- ui/ — Vite + React + TS + Tailwind + Motion + React Three FiberPaper & Ink editorial UI with 3D paper hero, dark/light mode,confidence inkwell, wax-stamp metrics, kinetic typography- 95 passing tests (up from 54); UI is a separate npm workspace 9 days ago
run_multimodel_benchmark.py

8.2 kB
eval: multi-model benchmark - nano is Pareto-optimal (0.896 micro F1 at 0.0116/doc) 7 days ago