KalanaPabasara commited on
Commit Β·
6c6f88b
1
Parent(s): 1fed70a
Add HF Spaces README and Streamlit config
Browse files- .streamlit/config.toml +11 -0
- README.md +67 -0
.streamlit/config.toml
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[theme]
|
| 2 |
+
primaryColor = "#FF4B4B"
|
| 3 |
+
backgroundColor = "#FFFFFF"
|
| 4 |
+
secondaryBackgroundColor = "#F0F2F6"
|
| 5 |
+
textColor = "#262730"
|
| 6 |
+
font = "sans serif"
|
| 7 |
+
|
| 8 |
+
[server]
|
| 9 |
+
headless = true
|
| 10 |
+
enableCORS = false
|
| 11 |
+
enableXsrfProtection = true
|
README.md
ADDED
|
@@ -0,0 +1,67 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
title: SinCode
|
| 3 |
+
emoji: π±π°
|
| 4 |
+
colorFrom: blue
|
| 5 |
+
colorTo: yellow
|
| 6 |
+
sdk: streamlit
|
| 7 |
+
sdk_version: 1.32.0
|
| 8 |
+
app_file: app.py
|
| 9 |
+
pinned: false
|
| 10 |
+
license: mit
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# ΰ·ΰ·ΰΆCode β Singlish to Sinhala Transliterator
|
| 14 |
+
|
| 15 |
+
A model-driven, context-aware back-transliteration system that converts Romanised Sinhala (Singlish) to native Sinhala script.
|
| 16 |
+
|
| 17 |
+
## Architecture (v3)
|
| 18 |
+
|
| 19 |
+
```
|
| 20 |
+
Input sentence
|
| 21 |
+
β
|
| 22 |
+
βΌ
|
| 23 |
+
Word Tokenizer
|
| 24 |
+
β
|
| 25 |
+
ββ Sinhala script? βββββββββββββββββββββββΊ Pass through unchanged
|
| 26 |
+
β
|
| 27 |
+
ββ English vocab (len β₯ 3)? ββββββββββββββΊ Pass through unchanged
|
| 28 |
+
β
|
| 29 |
+
ββ Singlish word?
|
| 30 |
+
β
|
| 31 |
+
βΌ
|
| 32 |
+
ByT5-small seq2seq
|
| 33 |
+
(top-5 candidates)
|
| 34 |
+
β
|
| 35 |
+
βΌ
|
| 36 |
+
XLM-RoBERTa MLM reranker
|
| 37 |
+
(contextual scoring)
|
| 38 |
+
β
|
| 39 |
+
βΌ
|
| 40 |
+
Best candidate
|
| 41 |
+
```
|
| 42 |
+
|
| 43 |
+
## Models
|
| 44 |
+
|
| 45 |
+
| Model | Role | Hub ID |
|
| 46 |
+
|-------|------|--------|
|
| 47 |
+
| ByT5-small | Singlish β Sinhala candidate generation | `Kalana001/byt5-small-singlish-sinhala` |
|
| 48 |
+
| XLM-RoBERTa | Contextual MLM reranking | `Kalana001/xlm-roberta-base-finetuned-sinhala` |
|
| 49 |
+
| mBart50 | Full-sentence Sinhala output mode | `Kalana001/mbart50-large-singlish-sinhala` |
|
| 50 |
+
|
| 51 |
+
## Modes
|
| 52 |
+
|
| 53 |
+
- **Code-Mixed Output** β Retains English words where contextually appropriate; Singlish words are transliterated using ByT5 + XLM-RoBERTa reranking.
|
| 54 |
+
- **Full Sinhala Output** β Transliterates the entire sentence to Sinhala script using mBart50.
|
| 55 |
+
|
| 56 |
+
## Environment Variables (optional)
|
| 57 |
+
|
| 58 |
+
Set these in HF Spaces β Settings β Repository secrets to enable Supabase feedback storage:
|
| 59 |
+
|
| 60 |
+
| Variable | Description |
|
| 61 |
+
|----------|-------------|
|
| 62 |
+
| `SUPABASE_URL` | Supabase project URL |
|
| 63 |
+
| `SUPABASE_ANON_KEY` | Supabase anon key |
|
| 64 |
+
| `SUPABASE_SERVICE_ROLE_KEY` | Supabase service role key |
|
| 65 |
+
| `SUPABASE_FEEDBACK_TABLE` | Table name (default: `feedback_submissions`) |
|
| 66 |
+
|
| 67 |
+
If not set, feedback is saved locally to `misc/feedback_submissions.jsonl`.
|