dream-cubed commited on
Commit
6cf2c51
·
verified ·
1 Parent(s): ad0daaf

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +32 -28
README.md CHANGED
@@ -1,28 +1,32 @@
1
- # Dream-Cubed Human Representative Sample
2
-
3
- This directory is a deterministic, class-stratified sample of the Dream-Cubed Human dataset. It is provided for reviewer inspection of data quality; the full dataset remains available at https://huggingface.co/datasets/dream-cubed/DreamCubedHumanSample.
4
-
5
- The sample dataset for procedurally generated data is available at https://huggingface.co/datasets/dream-cubed/DreamCubedNatural.
6
-
7
-
8
- Depending on which commands have been run, the sample may include raw human-authored map chunks and/or compact processed train/validation examples. Processed examples may also contain natural biome chunks from the companion Dream-Cubed Natural release. Permission to release voxelized chunks from the human-authored maps was granted explicitly by the anonymized map creator for noncommercial research, but not for commercial products or reconstruction of the original maps.
9
-
10
- ## Creation Procedure
11
-
12
- - Sampling seed: `67`
13
- - Last source mode added: `raw`
14
- - Raw target: up to `32` chunks per raw source label
15
- - Processed target: up to `128` chunks per label per split
16
- - Strategy: deterministic stratified sampling without replacement
17
- - Output: compact `.npz` raw samples and restored, unsharded processed `.npy` split samples
18
-
19
- Exact source paths, selected source indices, sample counts, and output files are recorded in
20
- `sample_manifest.json`. Processed split directories also include `source_indices.npy` and fresh
21
- `manifest.json` files. Dataset metadata files copied from nearby public dataset files are stored
22
- under `metadata/` when available.
23
-
24
- ## Inspection Notes
25
-
26
- This sample is intentionally small and should not be used for model training or quantitative
27
- evaluation. It is meant only to help reviewers inspect file formats, labels, chunk tensors, and
28
- overall data quality without downloading the full multi-GB dataset.
 
 
 
 
 
1
+ ---
2
+ license: cc-by-nc-4.0
3
+ ---
4
+
5
+ # Dream-Cubed Human Representative Sample
6
+
7
+ This directory is a deterministic, class-stratified sample of the Dream-Cubed Human dataset. It is provided for reviewer inspection of data quality; the full dataset remains available at https://huggingface.co/datasets/dream-cubed/DreamCubedHumanSample.
8
+
9
+ The sample dataset for procedurally generated data is available at https://huggingface.co/datasets/dream-cubed/DreamCubedNatural.
10
+
11
+
12
+ Depending on which commands have been run, the sample may include raw human-authored map chunks and/or compact processed train/validation examples. Processed examples may also contain natural biome chunks from the companion Dream-Cubed Natural release. Permission to release voxelized chunks from the human-authored maps was granted explicitly by the anonymized map creator for noncommercial research, but not for commercial products or reconstruction of the original maps.
13
+
14
+ ## Creation Procedure
15
+
16
+ - Sampling seed: `67`
17
+ - Last source mode added: `raw`
18
+ - Raw target: up to `32` chunks per raw source label
19
+ - Processed target: up to `128` chunks per label per split
20
+ - Strategy: deterministic stratified sampling without replacement
21
+ - Output: compact `.npz` raw samples and restored, unsharded processed `.npy` split samples
22
+
23
+ Exact source paths, selected source indices, sample counts, and output files are recorded in
24
+ `sample_manifest.json`. Processed split directories also include `source_indices.npy` and fresh
25
+ `manifest.json` files. Dataset metadata files copied from nearby public dataset files are stored
26
+ under `metadata/` when available.
27
+
28
+ ## Inspection Notes
29
+
30
+ This sample is intentionally small and should not be used for model training or quantitative
31
+ evaluation. It is meant only to help reviewers inspect file formats, labels, chunk tensors, and
32
+ overall data quality without downloading the full multi-GB dataset.