sahanruwantha / siconv

The siconv Python package simplifies the conversion between Sinhala and Singlish languages. Whether you're dealing with Sinhala text that needs transformation into Singlish or vice versa, siconv offers a user-friendly solution.
GNU General Public License v3.0
6 stars 2 forks source link

Increasing Accuracy of `sinhala_to_singlish` #25

Open IsiraUdaththa opened 1 month ago

IsiraUdaththa commented 1 month ago

As dicussed in #24 about sinhala_to_singlish. I have used the code from another repo. It gives me better accuracy with similar performance.

100%|████████████████████████| 440024/440024 [00:32<00:00, 13394.15it/s]

--- Test Results ---
Passed   : 423848
Failed   : 16176 
Accuracy : 0.963

I don't quite understand the all of the rules. I'll create a pull request in weekend.

IsiraUdaththa commented 1 month ago

I've rewrite the code and achieved even higher accuracy (99.7%) with additional tweaks. Given the significant changes, I'm unsure if I should commit these updates.

100%|██████████| 440024/440024 [00:34<00:00, 12753.35it/s]

--- Test Results ---
Passed   : 438908
Failed   : 1116
Accuracy : 0.997

You can review the updated code here. The tests function works as they did previously.

sahanruwantha commented 1 month ago

can you push the changes in a new branch, so could review it and repackage it.