Update README.md
Browse files
README.md
CHANGED
|
@@ -26,7 +26,6 @@ The Karibu project is a collaboration between pleIAs, Bibliothèque sans fronti
|
|
| 26 |
Karibu is a DeBERTa-based classifier that automatically assigns CEFR language proficiency levels (A1-C2) to French educational content.
|
| 27 |
|
| 28 |
## Architecture: DeBERTa with multi-head classification
|
| 29 |
-
- Base Model: PleIAs/celadon
|
| 30 |
- Model Size: Fine-tuned from DeBERTa-v3-small
|
| 31 |
- Output : 6 classification levels (A1, A2, B1, B2, C1, C2)
|
| 32 |
|
|
@@ -38,7 +37,7 @@ Karibu is a DeBERTa-based classifier that automatically assigns CEFR language pr
|
|
| 38 |
- Training Data: 9,000 synthetic samples
|
| 39 |
|
| 40 |
- Source: French press articles + Wikimedia content
|
| 41 |
-
- Processing: Sequential text simplification using an open source model (to
|
| 42 |
- Validation: 1,000 samples per level manually verified by BSF experts
|
| 43 |
|
| 44 |
## Topics Coverage:
|
|
|
|
| 26 |
Karibu is a DeBERTa-based classifier that automatically assigns CEFR language proficiency levels (A1-C2) to French educational content.
|
| 27 |
|
| 28 |
## Architecture: DeBERTa with multi-head classification
|
|
|
|
| 29 |
- Model Size: Fine-tuned from DeBERTa-v3-small
|
| 30 |
- Output : 6 classification levels (A1, A2, B1, B2, C1, C2)
|
| 31 |
|
|
|
|
| 37 |
- Training Data: 9,000 synthetic samples
|
| 38 |
|
| 39 |
- Source: French press articles + Wikimedia content
|
| 40 |
+
- Processing: Sequential text simplification using an open source model (to be published)
|
| 41 |
- Validation: 1,000 samples per level manually verified by BSF experts
|
| 42 |
|
| 43 |
## Topics Coverage:
|