Update README.md

README.md

@@ -137,11 +137,11 @@ The model was evaluated on a held-out test set of **5,081 samples**, covering a
 #### 1. Overall Performance
 | Metric | Score | Note |
 | :--- | :--- | :--- |
-| **BLEU** | **86.58** | High linguistic and semantic fidelity |
+| **BLEU** | **86.34** | High linguistic and semantic fidelity |
 | **Word Accuracy** | **93.63%** | Robust word-level correction |
 | **Exact Match** | **52.23%** | Entire sentence perfectly restored |
-| **WER** | **0.0897** | ~8.97% error rate per word |
-| **CER** | **0.0402** | ~4.02% error rate per character |
+| **WER** | **0.0838** | ~8.38% error rate per word |
+| **CER** | **0.0360** | ~3.60% error rate per character |
 
 *Note: The Exact Match score reflects the inherent ambiguity in the Vietnamese language (e.g., "muon" could be "muốn", "mượn", or "muộn"), where multiple correct interpretations may exist without broader paragraph context.*
 
@@ -150,9 +150,9 @@ The model's performance varies based on the complexity and length of the input:
 
 | Category | Length (words) | Accuracy | Sample Count |
 | :--- | :--- | :--- | :--- |
-| **Short** | < 10 | **61.40%** | 2,347 |
-| **Medium** | 10 - 30 | **47.47%** | 2,408 |
-| **Long** | > 30 | **21.47%** | 326 |
+| **Short** | < 10 | **60.88%** | 2,927 |
+| **Medium** | 10 - 30 | **47.83%** | 3,577 |
+| **Long** | > 30 | **25.91%** | 552 |
 
 *Analysis: The model performs exceptionally well on short to medium sentences. Accuracy declines on longer sequences (>30 words), likely due to the increased probability of cumulative errors and the 256-token limit.*