Update README.md

README.md

@@ -1,3 +1,42 @@
+---
+license: apache-2.0
+datasets:
+- polyglots/MADLAD_CulturaX_cleaned
+language:
+- si
+pipeline_tag: feature-extraction
+library_name: transformers
+tags:
+- tokenizer
+- SGPE
+- linguis_trie
+- token
+- tokenization
+- remeinium
+- transformer
+- linguistics
+- NLP
+- sinhala
+- BPE
+- GPE
+model-index:
+- name: SGPE-Sinhala
+  results:
+  - task:
+      type: feature-extraction
+    dataset:
+      name: MADLAD-400 (CulturaX Cleaned Sinhala subset)
+      type: polyglots/MADLAD_CulturaX_cleaned
+    metrics:
+    - name: Token-to-Word Ratio (TWR)
+      type: twr
+      value: 1.438
+      verified: false
+    - name: Characters per Token (CPT)
+      type: cpt
+      value: 4.48
+      verified: false
+---
 # Syllable is the Token: SGPE - Syllable-Aware Grapheme Pair Encoding
 
 **Remeinium Research**