Tom Aarsen commited on
Commit
2df654a
1 Parent(s): 314777e

Add Sentence Transformers support

Browse files
1_Pooling/config.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 1024,
3
+ "pooling_mode_cls_token": false,
4
+ "pooling_mode_mean_tokens": true,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false
9
+ }
README.md CHANGED
@@ -1,6 +1,8 @@
1
  ---
2
  tags:
3
  - mteb
 
 
4
  model-index:
5
  - name: multilingual-e5-large-instruct
6
  results:
@@ -5390,7 +5392,9 @@ This model has 24 layers and the embedding size is 1024.
5390
 
5391
  ## Usage
5392
 
5393
- Below is an example to encode queries and passages from the MS-MARCO passage ranking dataset.
 
 
5394
 
5395
  ```python
5396
  import torch.nn.functional as F
@@ -5433,6 +5437,36 @@ embeddings = average_pool(outputs.last_hidden_state, batch_dict['attention_mask'
5433
  embeddings = F.normalize(embeddings, p=2, dim=1)
5434
  scores = (embeddings[:2] @ embeddings[2:].T) * 100
5435
  print(scores.tolist())
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5436
  ```
5437
 
5438
  ## Supported Languages
 
1
  ---
2
  tags:
3
  - mteb
4
+ - sentence-transformers
5
+ - transformers
6
  model-index:
7
  - name: multilingual-e5-large-instruct
8
  results:
 
5392
 
5393
  ## Usage
5394
 
5395
+ Below are examples to encode queries and passages from the MS-MARCO passage ranking dataset.
5396
+
5397
+ ### Transformers
5398
 
5399
  ```python
5400
  import torch.nn.functional as F
 
5437
  embeddings = F.normalize(embeddings, p=2, dim=1)
5438
  scores = (embeddings[:2] @ embeddings[2:].T) * 100
5439
  print(scores.tolist())
5440
+ # => [[91.92852783203125, 67.580322265625], [70.3814468383789, 92.1330795288086]]
5441
+ ```
5442
+
5443
+ ### Sentence Transformers
5444
+
5445
+ ```python
5446
+ from sentence_transformers import SentenceTransformer
5447
+
5448
+ def get_detailed_instruct(task_description: str, query: str) -> str:
5449
+ return f'Instruct: {task_description}\nQuery: {query}'
5450
+
5451
+ # Each query must come with a one-sentence instruction that describes the task
5452
+ task = 'Given a web search query, retrieve relevant passages that answer the query'
5453
+ queries = [
5454
+ get_detailed_instruct(task, 'how much protein should a female eat'),
5455
+ get_detailed_instruct(task, '南瓜的家常做法')
5456
+ ]
5457
+ # No need to add instruction for retrieval documents
5458
+ documents = [
5459
+ "As a general guideline, the CDC's average requirement of protein for women ages 19 to 70 is 46 grams per day. But, as you can see from this chart, you'll need to increase that if you're expecting or training for a marathon. Check out the chart below to see how much protein you should be eating each day.",
5460
+ "1.清炒南瓜丝 原料:嫩南瓜半个 调料:葱、盐、白糖、鸡精 做法: 1、南瓜用刀薄薄的削去表面一层皮,用勺子刮去瓤 2、擦成细丝(没有擦菜板就用刀慢慢切成细丝) 3、锅烧热放油,入葱花煸出香味 4、入南瓜丝快速翻炒一分钟左右,放盐、一点白糖和鸡精调味出锅 2.香葱炒南瓜 原料:南瓜1只 调料:香葱、蒜末、橄榄油、盐 做法: 1、将南瓜去皮,切成片 2、油锅8成热后,将蒜末放入爆香 3、爆香后,将南瓜片放入,翻炒 4、在翻炒的同时,可以不时地往锅里加水,但不要太多 5、放入盐,炒匀 6、南瓜差不多软和绵了之后,就可以关火 7、撒入香葱,即可出锅"
5461
+ ]
5462
+ input_texts = queries + documents
5463
+
5464
+ model = SentenceTransformer('intfloat/multilingual-e5-large-instruct')
5465
+
5466
+ embeddings = model.encode(input_texts, convert_to_tensor=True, normalize_embeddings=True)
5467
+ scores = (embeddings[:2] @ embeddings[2:].T) * 100
5468
+ print(scores.tolist())
5469
+ # [[91.92853546142578, 67.5802993774414], [70.38143157958984, 92.13307189941406]]
5470
  ```
5471
 
5472
  ## Supported Languages
config_sentence_transformers.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "2.4.0.dev0",
4
+ "transformers": "4.37.0",
5
+ "pytorch": "2.1.0+cu121"
6
+ }
7
+ }
modules.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Normalize",
18
+ "type": "sentence_transformers.models.Normalize"
19
+ }
20
+ ]