Spaces:
Running
Running
Update README.md
Browse files
README.md
CHANGED
|
@@ -46,19 +46,6 @@ We adapt existing architectures, post-train them on focused tasks, and document
|
|
| 46 |
|
| 47 |
**Datasets:** [Tender Mermaid Training V1](https://huggingface.co/datasets/TendieLabs/Tender_Mermaid_Training_V1) and [V2](https://huggingface.co/datasets/TendieLabs/Tender_Mermaid_Training_V2), the in-house Mermaid corpora behind the Fred family.
|
| 48 |
|
| 49 |
-
## Turning Language Into Structure
|
| 50 |
-
|
| 51 |
-
The Fred family is built for one job: generating valid, renderable Mermaid syntax. Not approximate descriptions, but structured diagrams for databases, workflows, and STEM reasoning.
|
| 52 |
-
|
| 53 |
-
## How We Work
|
| 54 |
-
|
| 55 |
-
1. Start from a verified open base.
|
| 56 |
-
2. Apply targeted post-training.
|
| 57 |
-
3. Evaluate against specific capabilities.
|
| 58 |
-
4. Document the result in the model card.
|
| 59 |
-
|
| 60 |
-
We do not publish metrics that have not been independently verified. Training data and base models are linked and attributed in every model card.
|
| 61 |
-
|
| 62 |
## Good Research Starts With Curiosity.
|
| 63 |
|
| 64 |
We believe useful AI research should be explored openly, documented carefully, and made accessible wherever practical.
|
|
|
|
| 46 |
|
| 47 |
**Datasets:** [Tender Mermaid Training V1](https://huggingface.co/datasets/TendieLabs/Tender_Mermaid_Training_V1) and [V2](https://huggingface.co/datasets/TendieLabs/Tender_Mermaid_Training_V2), the in-house Mermaid corpora behind the Fred family.
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
## Good Research Starts With Curiosity.
|
| 50 |
|
| 51 |
We believe useful AI research should be explored openly, documented carefully, and made accessible wherever practical.
|