Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -23,7 +23,7 @@ Developed by **AMEFORGE**.
|
|
| 23 |
|
| 24 |
## Architecture & Parameters
|
| 25 |
- **Architecture:** SparseMind v15 (dynamic neural type management, sparse attention, and sparse FFN)
|
| 26 |
-
- **Parameters:**
|
| 27 |
- **Hidden dimension:** 384
|
| 28 |
- **Layers:** 8
|
| 29 |
- **Context length:** 512 tokens
|
|
@@ -32,7 +32,7 @@ Developed by **AMEFORGE**.
|
|
| 32 |
## Training Details
|
| 33 |
- Trained on **30,000,000+** procedurally generated programming templates on-the-fly.
|
| 34 |
- Masked loss to learn code completion and documentation formatting.
|
| 35 |
-
- Best Python syntax-validity on held-out validation:
|
| 36 |
|
| 37 |
## How to use
|
| 38 |
```python
|
|
|
|
| 23 |
|
| 24 |
## Architecture & Parameters
|
| 25 |
- **Architecture:** SparseMind v15 (dynamic neural type management, sparse attention, and sparse FFN)
|
| 26 |
+
- **Parameters:** 36,824,070
|
| 27 |
- **Hidden dimension:** 384
|
| 28 |
- **Layers:** 8
|
| 29 |
- **Context length:** 512 tokens
|
|
|
|
| 32 |
## Training Details
|
| 33 |
- Trained on **30,000,000+** procedurally generated programming templates on-the-fly.
|
| 34 |
- Masked loss to learn code completion and documentation formatting.
|
| 35 |
+
- Best Python syntax-validity on held-out validation: 8.8%
|
| 36 |
|
| 37 |
## How to use
|
| 38 |
```python
|