Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: cc0-1.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
# Jafar Checkpoint
|
| 6 |
+
|
| 7 |
+
Pretrained **MaskGIT-based world model** from the [Jafar](https://github.com/FLAIROx/jafar/) codebase.
|
| 8 |
+
Trained on the **CoinRun** dataset for action-conditioned video generation using the Jafar configuration as mentioned in the paper.
|
| 9 |
+
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
### Model Details
|
| 13 |
+
- **Architecture:** ST-Transformer (spatio-temporal transformer)
|
| 14 |
+
- **Input:** 16-frame sequences (64×64) + latent actions
|
| 15 |
+
- **Training Environment:** CoinRun (Cobbe et al., 2020)
|
| 16 |
+
- **Objective:** MaskGIT (Chang et al., 2022)
|