Datasets:
Modalities:
Image
Formats:
imagefolder
Languages:
English
Size:
< 1K
Tags:
Large Language Models
LLM Evaluation
Sequential Reasoning
Scaling Laws
Synthetic Benchmarks
Commonsense Reasoning
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -118,6 +118,28 @@ if instance_data:
|
|
| 118 |
# available in the main project repository.
|
| 119 |
```
|
| 120 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 121 |
## Dataset Statistics (for `seqBench_compact.jsonl.gz`)
|
| 122 |
|
| 123 |
* **Total Instances:** 7079
|
|
|
|
| 118 |
# available in the main project repository.
|
| 119 |
```
|
| 120 |
|
| 121 |
+
## Maze Visualization
|
| 122 |
+
|
| 123 |
+
### To Create Maze Images
|
| 124 |
+
|
| 125 |
+
Use the **`maze_visualizer_utility.py`** in scripts folder to generate maze visualizations from SeqBench dataset instances.
|
| 126 |
+
This utility supports several visualization modes and automatically generates PNG files with descriptive names.
|
| 127 |
+
|
| 128 |
+
**Quick Examples:**
|
| 129 |
+
```bash
|
| 130 |
+
# Empty maze structure only
|
| 131 |
+
python maze_visualizer_utility.py --maze_id <instance_id> --mode empty
|
| 132 |
+
|
| 133 |
+
# Optimal path visualization
|
| 134 |
+
python maze_visualizer_utility.py --maze_id <instance_id> --mode optimal
|
| 135 |
+
|
| 136 |
+
# Model solution only
|
| 137 |
+
python maze_visualizer_utility.py --maze_id <instance_id> --result_file results/path/to/91_run3.json --mode model
|
| 138 |
+
|
| 139 |
+
# Side-by-side comparison (optimal vs model solution)
|
| 140 |
+
python maze_visualizer_utility.py --maze_id <instance_id> --result_file results/path/to/result.json
|
| 141 |
+
```
|
| 142 |
+
|
| 143 |
## Dataset Statistics (for `seqBench_compact.jsonl.gz`)
|
| 144 |
|
| 145 |
* **Total Instances:** 7079
|