Pinned Loading

1. Maze_RL Maze_RL Public

   A custom POMDP maze environment for studying agent reasoning, uncertainty, partial observation, and world-model learning. Produced structured trajectories for training GPT-based reasoning models.

   Python 2

2. PlainGPT PlainGPT Public

   A compact, fully self-contained Transformer framework for LoRA fine-tuning and text generation.

   Python 1

3. AntWorld AntWorld Public

   AntWorld – Multi-ant grid simulation with local memory, pheromones, and 5×5 communication（多螞蟻網格模擬，用於世界模型與強化學習實驗）

   Python 1

4. QT-R1 QT-R1 Public

   STaR × S1 math pipeline on Qwen2.5-1.5B. LoRA, strict Final: format, ~20–30% acc (OpenR1-Math split).

   Python 1