Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -39,6 +39,7 @@ CTM's key innovation: **accuracy improves with more internal iterations**. The m
|
|
| 39 |
| Tracking-Quadrant | `ctm-tracking-quadrant.pt` | 6.7M | Motion quadrant | 100% | 4-class prediction |
|
| 40 |
| Tracking-Position | `ctm-tracking-position.pt` | 6.7M | Exact position | 93.8% | 256-class (16x16 grid) |
|
| 41 |
| Transfer | `ctm-transfer-parity-brackets.pt` | 2.5M | Transfer learning | 94.5% | Parity core to brackets |
|
|
|
|
| 42 |
|
| 43 |
## Model Configurations
|
| 44 |
|
|
@@ -112,6 +113,22 @@ config = {
|
|
| 112 |
}
|
| 113 |
```
|
| 114 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 115 |
## Usage
|
| 116 |
|
| 117 |
```python
|
|
|
|
| 39 |
| Tracking-Quadrant | `ctm-tracking-quadrant.pt` | 6.7M | Motion quadrant | 100% | 4-class prediction |
|
| 40 |
| Tracking-Position | `ctm-tracking-position.pt` | 6.7M | Exact position | 93.8% | 256-class (16x16 grid) |
|
| 41 |
| Transfer | `ctm-transfer-parity-brackets.pt` | 2.5M | Transfer learning | 94.5% | Parity core to brackets |
|
| 42 |
+
| Jigsaw MNIST | `ctm-jigsaw-mnist.pt` | 19M | Jigsaw puzzle solving | 92.3% | Reassemble 2x2 shuffled MNIST |
|
| 43 |
|
| 44 |
## Model Configurations
|
| 45 |
|
|
|
|
| 113 |
}
|
| 114 |
```
|
| 115 |
|
| 116 |
+
### Jigsaw MNIST CTM
|
| 117 |
+
```python
|
| 118 |
+
config = {
|
| 119 |
+
"iterations": 30,
|
| 120 |
+
"memory_length": 20,
|
| 121 |
+
"d_model": 512,
|
| 122 |
+
"d_input": 128,
|
| 123 |
+
"heads": 8,
|
| 124 |
+
"n_synch_out": 32,
|
| 125 |
+
"n_synch_action": 32,
|
| 126 |
+
"synapse_depth": 1,
|
| 127 |
+
"out_dims": 24, # 4 tiles x 6 permutation options
|
| 128 |
+
"backbone_type": "jigsaw",
|
| 129 |
+
}
|
| 130 |
+
```
|
| 131 |
+
|
| 132 |
## Usage
|
| 133 |
|
| 134 |
```python
|