nreimers commited on
Commit
a088dbe
1 Parent(s): c40a747

update readme

Browse files
Files changed (1) hide show
  1. README.md +9 -9
README.md CHANGED
@@ -1335,13 +1335,12 @@ int8 embeddings keep 99.7-100% of the search quality, while reducing your vector
1335
  Below example using [usearch](https://github.com/unum-cloud/usearch) to search on int8 embeddings.
1336
 
1337
  ```python
1338
- #Run: pip install cohere datasets numpy usearch
1339
  from datasets import load_dataset
1340
  import numpy as np
1341
  import cohere
1342
  from usearch.index import Index
1343
 
1344
- co = cohere.Client("YOUR_COHERE_API_KEY") # Add your cohere API key from www.cohere.com
1345
  lang = "simple"
1346
 
1347
  #Load at max 1000 chunks + embeddings
@@ -1352,11 +1351,13 @@ docs = []
1352
  doc_embeddings = []
1353
  for doc in docs_stream:
1354
  docs.append(doc)
1355
- doc_embeddings.append(doc['emb_ubinary'])
1356
  if len(docs) >= max_docs:
1357
  break
1358
 
1359
- doc_embeddings = np.asarray(doc_embeddings)
 
 
1360
 
1361
  #Create the usearch HNSW index on the int8 embeddings
1362
  num_dim = 1024
@@ -1374,9 +1375,8 @@ matches = index.search(query_emb, 10)
1374
  # Print results
1375
  for match in matches:
1376
  doc_id = match.key
1377
- row = df.iloc[doc_id]
1378
- print(row['title'])
1379
- print(row['text'])
1380
  print("----")
1381
  ```
1382
 
@@ -1392,7 +1392,7 @@ import numpy as np
1392
  import cohere
1393
  import faiss
1394
 
1395
- co = cohere.Client("YOUR_COHERE_API_KEY") # Add your cohere API key from www.cohere.com
1396
  lang = "simple"
1397
 
1398
  #Load at max 1000 chunks + embeddings
@@ -1407,7 +1407,7 @@ for doc in docs_stream:
1407
  if len(docs) >= max_docs:
1408
  break
1409
 
1410
- doc_embeddings = np.asarray(doc_embeddings)
1411
 
1412
  #Create the faiss IndexBinaryFlat index
1413
  num_dim = 1024
 
1335
  Below example using [usearch](https://github.com/unum-cloud/usearch) to search on int8 embeddings.
1336
 
1337
  ```python
 
1338
  from datasets import load_dataset
1339
  import numpy as np
1340
  import cohere
1341
  from usearch.index import Index
1342
 
1343
+ co = cohere.Client("<<YOUR_API_KEY>>") # Add your cohere API key from www.cohere.com
1344
  lang = "simple"
1345
 
1346
  #Load at max 1000 chunks + embeddings
 
1351
  doc_embeddings = []
1352
  for doc in docs_stream:
1353
  docs.append(doc)
1354
+ doc_embeddings.append(doc['emb_int8'])
1355
  if len(docs) >= max_docs:
1356
  break
1357
 
1358
+ doc_embeddings = np.asarray(doc_embeddings, dtype='int8')
1359
+
1360
+ print(doc_embeddings.shape, doc_embeddings.dtype)
1361
 
1362
  #Create the usearch HNSW index on the int8 embeddings
1363
  num_dim = 1024
 
1375
  # Print results
1376
  for match in matches:
1377
  doc_id = match.key
1378
+ print(docs[doc_id]['title'])
1379
+ print(docs[doc_id]['text'])
 
1380
  print("----")
1381
  ```
1382
 
 
1392
  import cohere
1393
  import faiss
1394
 
1395
+ co = cohere.Client("<<YOUR_API_KEY>>") # Add your cohere API key from www.cohere.com
1396
  lang = "simple"
1397
 
1398
  #Load at max 1000 chunks + embeddings
 
1407
  if len(docs) >= max_docs:
1408
  break
1409
 
1410
+ doc_embeddings = np.asarray(doc_embeddings, dtype='uint8')
1411
 
1412
  #Create the faiss IndexBinaryFlat index
1413
  num_dim = 1024