Link externally hosted checkpoint weight
Browse files
README.md
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc-by-nc-4.0
|
| 3 |
+
tags: [bagel, vlm-gym, sokoban, reinforcement-learning]
|
| 4 |
+
---
|
| 5 |
+
|
| 6 |
+
# sokoban_adaptive_vanilla3k_rl2box_step100
|
| 7 |
+
|
| 8 |
+
The model card remains in the original UWM namespace. The BF16 checkpoint weight is hosted in
|
| 9 |
+
[`JiaxinGe/umw`](https://huggingface.co/datasets/JiaxinGe/umw/tree/main/checkpoints/sokoban_adaptive_vanilla3k_rl2box_step100) because this namespace reached its public LFS storage quota.
|
| 10 |
+
|
| 11 |
+
- [Download `ema.safetensors`](https://huggingface.co/datasets/JiaxinGe/umw/resolve/main/checkpoints/sokoban_adaptive_vanilla3k_rl2box_step100/ema.safetensors?download=true)
|
| 12 |
+
- [Evaluation trajectories](https://huggingface.co/datasets/novastar111/uwm_paper_eval_trajectories)
|