| license: mit | |
| task_categories: | |
| - reinforcement-learning | |
| tags: | |
| - gymnasium | |
| - lunar-lander | |
| - imitation-learning | |
| - embeddings | |
| - mobilenet-v3 | |
| # Mind Games — Features | |
| Pre-computed visual features for the [mind-games](https://github.com/kvark/mind-games) project. | |
| ## Lunar Lander (MobileNetV3-Small) | |
| 200 episodes of MobileNetV3-Small embeddings extracted from LunarLander-v3 gameplay frames. | |
| - **Backbone**: MobileNetV3-Small (ImageNet pretrained, frozen) | |
| - **Embedding dim**: 576 (global average pooled) | |
| - **Precision**: float16 | |
| - **Size**: ~108MB | |
| - **Structure**: `lunar_lander/episode_NNNN/` with `embeddings.npy` (N×576) and `actions.npy` (N,) | |
| - **Index**: `lunar_lander/index.json` with per-episode reward and step count | |
| These embeddings enable fast CPU-only training (~4s/epoch vs ~67s end-to-end). | |