johnlockejrr
/

marianmt_syr_voc_eastern

Model card Files Files and versions

johnlockejrr commited on 8 days ago

Commit

39d03e9

·

verified ·

1 Parent(s): fa65c54

Update README.md

Files changed (1) hide show

README.md +4 -4

README.md CHANGED Viewed

@@ -42,7 +42,7 @@ model-index:
 # MarianMT Eastern Syriac Vocalization Model
-A fine-tuned MarianMT model for automatic Eastern Syriac (Mossul dialect) vocalization, converting consonantal (unvocalized) Syriac text to fully vocalized text with diacritical marks.
 ## Model Description
@@ -51,9 +51,9 @@ This model is fine-tuned from [`Helsinki-NLP/opus-mt-tc-bible-big-sem-en`](https
 ### Key Features
 - **Single-direction model**: Converts consonantal Syriac (`>>syr_cons<<`) to vocalized Eastern Syriac (`>>syr_voc<<`)
-- **Eastern Syriac optimized**: Trained specifically on Eastern Syriac texts (Mossul dialect) and Digital Syriac Corpus texts vocalized in Eastern Syriac
 - **High performance**: Achieves 62.41 BLEU, 87.98 chrF, and 58.81% character accuracy on test set
-- **Biblical and corpus text optimized**: Trained on Eastern Syriac Bible texts (Mossul) and Digital Syriac Corpus texts
 ## Model Details
@@ -181,7 +181,7 @@ Recommended generation parameters:
 ## Limitations and Bias
-- **Dialect Specificity**: This model is trained specifically on Eastern Syriac (Mossul dialect). Performance may vary on Western Syriac or other Syriac dialects.
 - **Domain Specificity**: This model is trained primarily on biblical and corpus Syriac texts. Performance may vary on other domains (e.g., modern Syriac, poetry, prose).
 - **Single Direction**: The model only vocalizes consonantal text. It does not perform the reverse operation (removing vocalization).
 - **Length Constraints**: Maximum input/output length is 512 tokens. Longer texts should be split into smaller segments.

 # MarianMT Eastern Syriac Vocalization Model
+A fine-tuned MarianMT model for automatic Eastern Syriac (Mossul Bible) vocalization, converting consonantal (unvocalized) Syriac text to fully vocalized text with diacritical marks.
 ## Model Description
 ### Key Features
 - **Single-direction model**: Converts consonantal Syriac (`>>syr_cons<<`) to vocalized Eastern Syriac (`>>syr_voc<<`)
+- **Eastern Syriac optimized**: Trained specifically on Eastern Syriac texts (Mossul edition) and Digital Syriac Corpus texts vocalized in Eastern Syriac
 - **High performance**: Achieves 62.41 BLEU, 87.98 chrF, and 58.81% character accuracy on test set
+- **Biblical and corpus text optimized**: Trained on Eastern Syriac Bible texts (Mossul edition) and Digital Syriac Corpus texts
 ## Model Details
 ## Limitations and Bias
+- **Dialect Specificity**: This model is trained specifically on Eastern Syriac (Mossul edition). Performance may vary on Western Syriac or other Syriac dialects.
 - **Domain Specificity**: This model is trained primarily on biblical and corpus Syriac texts. Performance may vary on other domains (e.g., modern Syriac, poetry, prose).
 - **Single Direction**: The model only vocalizes consonantal text. It does not perform the reverse operation (removing vocalization).
 - **Length Constraints**: Maximum input/output length is 512 tokens. Longer texts should be split into smaller segments.