Hugging Face's logo

Datasets:
cryptolab-playground
/
gas-centroids

Tasks:
Text Retrieval
Image-to-Text
Sentence Similarity
Languages:
English
Tags:
embeddings
vector-database
benchmark
Dataset card Files
xet
Community
4
suyeong-park commited on
Commit
2b9053e
·
verified ·
1 Parent(s): cb74231

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +4 -1
README.md CHANGED
@@ -33,14 +33,17 @@ For each embedding model, the directory contains two key file:
33
 
34
  | Data | Description |
35
  |-------|-------------|
36
- | `centroids.npy` | 32,768 centroids as followed IVF |
37
  | `tree_info.pkl` | tree metadata with parent and leaf info |
38
 
39
  ## Data Fields
40
 
41
  - Centroids: `centroids.npy`
 
42
  - numpy array(`np.ndarray`) with 768-dimensional 32,768 centroids
 
43
  - Tree Metadata: `tree_info.pkl`
 
44
  - keys: node_parents, leaf_ids, and leaf_to_centroid_idx
45
  - Detailed parsing information is provided in `es2`
46
 
 
33
 
34
  | Data | Description |
35
  |-------|-------------|
36
+ | `centroids.npy` | centroids as followed IVF |
37
  | `tree_info.pkl` | tree metadata with parent and leaf info |
38
 
39
  ## Data Fields
40
 
41
  - Centroids: `centroids.npy`
42
+ - This data is for finding the nearest clusters for IVF
43
  - numpy array(`np.ndarray`) with 768-dimensional 32,768 centroids
44
+
45
  - Tree Metadata: `tree_info.pkl`
46
+ - This data is for finding the virtual cluster following tree structure
47
  - keys: node_parents, leaf_ids, and leaf_to_centroid_idx
48
  - Detailed parsing information is provided in `es2`
49