BayanSimplify-ONNX
The on-device bundles of the BayanSimplify models: what the Bayan Android app downloads. Each folder is one bundle:
encoder.onnx and decoder.onnx (int8 dynamic quantization, merged decoder with KV cache), tokenizer.json, and
bayan_model.json (prefix, decoding settings, the decoder's inputs and outputs). The app checks every file's SHA-256
before using it.
| Folder | Model | In the app | Size |
|---|---|---|---|
v0.3/ |
BayanSimplify-v0.3, our best, decoded with four beams | "Large model" | 471 MB |
v0.2-Fast/ |
BayanSimplify-v0.2-Fast (AraBART) | "Fast model", the default download | 222 MB |
v0.2/ |
BayanSimplify-v0.2 (AraT5v2) | "Earlier large model" | 471 MB |
model3/, arabart/ and arat5/ hold the same bundles for app builds from before 5 October and will be removed.
In the app, each output passes through its text step: scripture and set poetry are never rewritten, and a sentence keeps its source wording when a rewrite drops a number or a Latin-script word, or changes a negation, limit or condition word.
The BayanSimplify family
| Model | What it is | Base model | Training data | In the Bayan app |
|---|---|---|---|---|
| BayanSimplify-v0.1 | Model 1 | AraT5v2-base-1024 | SAMER, level 5 → 3 | not shipped |
| BayanSimplify-v0.2 | Model 2 | AraT5v2-base-1024 | 22,733-pair mix (SAMER, Baseet, DAASI) with strength tags | "Earlier large model" |
| BayanSimplify-v0.2-Fast | Model 2, compact | AraBART | the same mix | "Fast model", the default download |
| BayanSimplify-v0.3 | Model 3, our best | AraT5v2-base-1024 | Bayan corpus v1, 14,975 rows | "Large model", four beams |
| BayanSimplify-ONNX | The int8 bundles the app downloads | v0.2-Fast, v0.2, v0.3 |
Licence and data
CC BY-NC 4.0 (non-commercial), in line with the training data's terms. SAMER is used under the CAMeL Lab's permission to fine-tune and share weights for non-commercial use; its text is not redistributed here.
About Bayan
Bayan (بيان) simplifies Arabic text for readers with dyslexia, entirely on an Android phone: select text in any app, choose «تبسيط» (Simplify), and a simpler version appears over the page. Built by Team Cogni for the Samsung Innovation Campus AI capstone, 2026. Every number on this card comes from the team's final report and BayanBench v2.0, the benchmark built for the task (meaning scored by Gemma 4 31B, checked against human raters, AUC 0.85).