Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,40 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: bsd-3-clause
|
| 3 |
+
language: ar
|
| 4 |
+
inference: false
|
| 5 |
+
tags:
|
| 6 |
+
- secryst
|
| 7 |
+
- transliteration
|
| 8 |
+
- diacritization
|
| 9 |
+
- onnx
|
| 10 |
+
- byt5
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
# ara-diac-small-2.1
|
| 14 |
+
|
| 15 |
+
Part of the [Secryst](https://secryst.github.io) phonological-layer family — a dedicated model for **diacritization** (AR), distributed as a Secryst IMF artifact and loadable through any Secryst crystal:
|
| 16 |
+
|
| 17 |
+
```python
|
| 18 |
+
from secryst import Model
|
| 19 |
+
Model.load("ara-diac-small-2.1")
|
| 20 |
+
```
|
| 21 |
+
|
| 22 |
+
- Artifact: `ara-diac-small-2.1-fp32.zip` (1.32 GiB)
|
| 23 |
+
- Precision variants: `ara-diac-small-2.1-int8`, `ara-diac-small-2.1-fp16`
|
| 24 |
+
|
| 25 |
+
## Measured quality
|
| 26 |
+
|
| 27 |
+
Full evaluation sets, zero skipped examples, the benchmark's own evaluator. Provenance for every number is the results record of the training project.
|
| 28 |
+
|
| 29 |
+
| metric | value | provenance |
|
| 30 |
+
|---|---|---|
|
| 31 |
+
| der_teacher_fullset | 2.289 | interscript/interscript-ml docs/RESULTS.md#run-006-r7-muon |
|
| 32 |
+
| der_student_fullset | 4.5701 | interscript/interscript-ml docs/RESULTS.md#run-006-r7-muon |
|
| 33 |
+
|
| 34 |
+
## License
|
| 35 |
+
|
| 36 |
+
BSD-3-CLAUSE
|
| 37 |
+
|
| 38 |
+
## Provenance
|
| 39 |
+
|
| 40 |
+
Trained in the [Secryst training monorepo](https://github.com/secryst/secryst-train) and published through the [interscript-ml](https://github.com/interscript/interscript-ml) distribution contract (models.yaml index, sha256-verified release assets, parity-gated artifacts). The canonical download channel is GitHub Releases; this repository is the Hugging Face mirror of the same bytes, pinned to the index revision.
|