Update README.md
Browse files
README.md
CHANGED
|
@@ -36,6 +36,16 @@ Each split is sharded:
|
|
| 36 |
- `states_{shard}.bin` — `np.float32` states (see below)
|
| 37 |
- `metadata.json` / `metadata_{shard}.json` — overall vs. per‐shard metadata
|
| 38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
### State Index Definition (New)
|
| 40 |
```
|
| 41 |
0: HIP_YAW
|
|
@@ -65,15 +75,6 @@ Each split is sharded:
|
|
| 65 |
24: Angular Velocity
|
| 66 |
```
|
| 67 |
|
| 68 |
-
---
|
| 69 |
-
|
| 70 |
-
## Test v2.0
|
| 71 |
-
|
| 72 |
-
We provide a 450 sample **test_v2.0** dataset for the [World Model Compression Challenge](https://huggingface.co/spaces/1x-technologies/1X_World_Model_Challenge_Compression) with a similar structure (`video_{shard}.bin`, `states_{shard}.bin`). Use:
|
| 73 |
-
- `unpack_data_test.py` to read the test set
|
| 74 |
-
- `unpack_data_train_val.py` to read train/val
|
| 75 |
-
---
|
| 76 |
-
|
| 77 |
## Previous v1.1
|
| 78 |
|
| 79 |
- `video.bin` — 16×16 patches at 30Hz, quantized
|
|
|
|
| 36 |
- `states_{shard}.bin` — `np.float32` states (see below)
|
| 37 |
- `metadata.json` / `metadata_{shard}.json` — overall vs. per‐shard metadata
|
| 38 |
|
| 39 |
+
|
| 40 |
+
---
|
| 41 |
+
|
| 42 |
+
## Test v2.0
|
| 43 |
+
|
| 44 |
+
We provide a 450 sample **test_v2.0** dataset for the [World Model Compression Challenge](https://huggingface.co/spaces/1x-technologies/1X_World_Model_Challenge_Compression) with a similar structure (`video_{shard}.bin`, `states_{shard}.bin`). Use:
|
| 45 |
+
- `unpack_data_test.py` to read the test set
|
| 46 |
+
- `unpack_data_train_val.py` to read train/val
|
| 47 |
+
---
|
| 48 |
+
|
| 49 |
### State Index Definition (New)
|
| 50 |
```
|
| 51 |
0: HIP_YAW
|
|
|
|
| 75 |
24: Angular Velocity
|
| 76 |
```
|
| 77 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
## Previous v1.1
|
| 79 |
|
| 80 |
- `video.bin` — 16×16 patches at 30Hz, quantized
|