Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -14,6 +14,7 @@ This repository contains FAISS index files created with the following parameters
|
|
| 14 |
- **specter2_IVF100,PQ16.index**: FAISS index file
|
| 15 |
- **specter2_IVF100,PQ64.index**: FAISS index file
|
| 16 |
- **emb_specter2.memmap**: Embedding memmap file
|
|
|
|
| 17 |
|
| 18 |
## Usage
|
| 19 |
|
|
@@ -36,6 +37,10 @@ embeddings = np.memmap(
|
|
| 36 |
dtype=np.float32
|
| 37 |
).reshape(-1, 768) # Adjust shape as needed
|
| 38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
# Example query
|
| 40 |
from transformers import AutoTokenizer, AutoModel
|
| 41 |
import torch
|
|
|
|
| 14 |
- **specter2_IVF100,PQ16.index**: FAISS index file
|
| 15 |
- **specter2_IVF100,PQ64.index**: FAISS index file
|
| 16 |
- **emb_specter2.memmap**: Embedding memmap file
|
| 17 |
+
- **patents_all.parquet**: Corpus parquet file
|
| 18 |
|
| 19 |
## Usage
|
| 20 |
|
|
|
|
| 37 |
dtype=np.float32
|
| 38 |
).reshape(-1, 768) # Adjust shape as needed
|
| 39 |
|
| 40 |
+
# Load corpus
|
| 41 |
+
import pandas as pd
|
| 42 |
+
corpus = pd.read_parquet("path/to/downloaded/corpus.parquet")
|
| 43 |
+
|
| 44 |
# Example query
|
| 45 |
from transformers import AutoTokenizer, AutoModel
|
| 46 |
import torch
|