Reproducing-TRM
Collection
3 items
•
Updated
•
4
TRM Model for Maze Solving
Model Description
This is a Tiny Recursive Model (TRM) fine-tuned for solving maze navigation tasks. The model implements recursive reasoning to find paths in 30x30 grid mazes.
- Developed by: alphaXiv
- Model type: TRM-Attention
- Language(s) (NLP): N/A (grid-based reasoning)
- License: MIT
- Finetuned from model: Custom TRM architecture
Intended Use
Primary Use
This model is designed to solve maze pathfinding problems by predicting the correct sequence of moves to navigate from start to goal in grid-based mazes.
Out-of-Scope Use
Not intended for general NLP tasks, image classification, or other domains outside maze solving.
Limitations and Bias
- Trained only on synthetic maze data
- May not generalize to mazes of different sizes or complexities
- Performance may degrade on mazes with unusual patterns
Training Data
The model was trained on a dataset of 30x30 grid mazes with hard difficulty levels. The dataset includes:
- Start and goal positions
- Wall configurations
- Correct path sequences
Evaluation Results
| Metric | Claimed | Achieved |
|---|---|---|
| Exact Accuracy | 85.3% | 83.67% ± 2.28% |
Results from independent reproduction study.
Repository
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support