ner

.gitattributes

@@ -58,4 +58,3 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
-qa/train.csv filter=lfs diff=lfs merge=lfs -text

README.md

@@ -33,6 +33,7 @@ tags:
 - unit-conversion
 - math
 - multilingual
+- persian
 - synthetic
 configs:
 - config_name: units
@@ -65,6 +66,11 @@ A three-part dataset about unit conversion:
 
 Every number is computed by code with exact rational arithmetic (`fractions.Fraction`) and rounded to 6 significant digits. No language model produced any answer. The generator is included (`build_dataset.py`, `langdata_names.py`, `langdata_templates.py`), so the whole dataset is reproducible (seed `20261105`).
 
+## Why this dataset
+
+* Language models still make mistakes on unit arithmetic, especially on the traps below.
+* Persian and Arabic resources for this domain are very scarce.
+* It includes **traditional Persian units** (mesqal, sir, man, zar, farsang, jarib) next to SI, US and imperial units.
 
 ## Loading
 
@@ -178,7 +184,7 @@ Each row has `text` and `entities`, a list of `{start, end, text, value, unit_id
 
 CC0 1.0 (public domain dedication). Sources of constants: BIPM SI Brochure, NIST SP 811 and Handbook 44, IEC 80000-13.
 
-[https://toolshed.click/](https://toolshed.click/)
+
 ---
 license: cc0-1.0
----
+---

qa/train.csv

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ebf33daf94e8d047d4231ddafdfcb941a78732bd1b148ec868487ee4d4c77526
-size 11844293