lukawskikacper commited on
Commit
5e060c3
1 Parent(s): f9c4cb7

Add README

Browse files
Files changed (2) hide show
  1. README.md +25 -0
  2. arxiv-abstracts-instructorxl.parquet +2 -2
README.md ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # arxiv-abstracts-instructorxl
2
+
3
+ This dataset contains 768-dimensional embeddings generated from the [arxiv](https://arxiv.org/)
4
+ paper abstracts using [InstructorXL](https://huggingface.co/hkunlp/instructor-xl) model. Each
5
+ vector has an abstract used to create it, along with the DOI (Digital Object Identifier). The
6
+ dataset was created using precomputed embeddings exposed by the [Alexandria Index](https://alex.macrocosm.so/download).
7
+
8
+ ## Generation process
9
+
10
+ The embeddings have been generated using the following instruction:
11
+
12
+ ```text
13
+ Represent the Research Paper abstract for retrieval; Input:
14
+ ```
15
+
16
+ The following code snippet shows how to generate embeddings using the InstructorXL model:
17
+
18
+ ```python
19
+ from InstructorEmbedding import INSTRUCTOR
20
+
21
+ model = INSTRUCTOR('hkunlp/instructor-xl')
22
+ sentence = "The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train."
23
+ instruction = "Represent the Research Paper abstract for retrieval; Input:"
24
+ embeddings = model.encode([[instruction, sentence]])
25
+ ```
arxiv-abstracts-instructorxl.parquet CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:cf9ac01969d810158ef0503757c6fe3d86bee9eeb37787d1e5d35919bffce17c
3
- size 16757338861
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ed5ca424c1ae92dec74b7b2ba96155a8bc55111c9b3c0ab4af9372c84f566dcf
3
+ size 16796999086