Upload kashmiri_char_tokenizer/README.md with huggingface_hub

kashmiri_char_tokenizer/README.md

@@ -23,7 +23,7 @@ datasets:
 | Architecture | Character-Level |
 | Language | Kashmiri (ks / kas) |
 | Script | Perso-Arabic (Nastaliq) |
-| Vocabulary Size | 135 |
+| Vocabulary Size | 134 |
 | Training Corpus | KS-LIT-3M (3,091,180 words) |
 | License | Apache-2.0 |
 
@@ -33,9 +33,9 @@ datasets:
 |--------|-------|-------------|
 | Fertility | 5.2453 | Tokens per word (lower = better) |
 | Diacritic Preservation Score (DPS) | 0.0000 | Novel KS-specific metric (1.0 = perfect) |
-| Morphological Boundary Alignment (MBA) | 0.2434 | IoU with gold morpheme boundaries |
+| Morphological Boundary Alignment (MBA) | 0.1994 | IoU with gold morpheme boundaries |
 | OOV Rate (held-out) | 0.0000 | Tested on unseen text |
-| Composite Quality Score (CQS) | 0.2595 | Weighted combination |
+| Composite Quality Score (CQS) | 0.2895 | Weighted combination |
 
 ## 🎯 Recommended Use Cases