Upload Maaza-SLM-360M-JSON-v1 - v1.0.0 production release
Browse files
README.md
CHANGED
|
@@ -114,7 +114,7 @@ Comparison to MLM-135M demonstrates scaling effectiveness:
|
|
| 114 |
## Training Data
|
| 115 |
|
| 116 |
### Dataset: EdgeJSON v3
|
| 117 |
-
- **Total Examples**: 787 (
|
| 118 |
- **Train Split**: 629 examples (80%)
|
| 119 |
- **Test Split**: 158 examples (20%)
|
| 120 |
- **Validation Rate**: 100% (all examples pass schema validation)
|
|
@@ -339,7 +339,7 @@ If you use this model in your research, please cite:
|
|
| 339 |
|
| 340 |
### v1.0.0 (2025-11-20)
|
| 341 |
- Initial release
|
| 342 |
-
- Trained on EdgeJSON v3 dataset (
|
| 343 |
- 55.1% JSONExact, 0.729 Field F1
|
| 344 |
- LoRA fine-tuning (r=32, alpha=64)
|
| 345 |
- 90.1 second training time
|
|
|
|
| 114 |
## Training Data
|
| 115 |
|
| 116 |
### Dataset: EdgeJSON v3
|
| 117 |
+
- **Total Examples**: 787 (validated)
|
| 118 |
- **Train Split**: 629 examples (80%)
|
| 119 |
- **Test Split**: 158 examples (20%)
|
| 120 |
- **Validation Rate**: 100% (all examples pass schema validation)
|
|
|
|
| 339 |
|
| 340 |
### v1.0.0 (2025-11-20)
|
| 341 |
- Initial release
|
| 342 |
+
- Trained on EdgeJSON v3 dataset (validated)
|
| 343 |
- 55.1% JSONExact, 0.729 Field F1
|
| 344 |
- LoRA fine-tuning (r=32, alpha=64)
|
| 345 |
- 90.1 second training time
|