yulongl commited on
Commit
472b2b9
1 Parent(s): 6d8105b

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +10 -10
README.md CHANGED
@@ -10,19 +10,19 @@ The detailed implementation of the model can be found in:
10
  https://arxiv.org/pdf/2112.08185.pdf
11
 
12
  Source code to train the model can be found via PrimeQA's IR component:
13
- https://github.com/primeqa/primeqa/tree/updated-documentation-readme/primeqa/ir/dense/colbert_top
14
 
15
  It is a Neural IR model built on top of the ColBERTv1 api and not directly compatible with Huggingface API. The inference result on XOR Dev dataset is:
16
  ```
17
- R@2kt R@5kt
18
- te 79.41 83.19
19
- bn 77.96 82.89
20
- fi 65.92 72.61
21
- ja 63.07 67.63
22
- ko 60.35 68.07
23
- ru 60.76 68.35
24
- ar 65.70 73.14
25
- Avg 67.60 73.70
26
  ```
27
 
28
  # Limitations and Bias
 
10
  https://arxiv.org/pdf/2112.08185.pdf
11
 
12
  Source code to train the model can be found via PrimeQA's IR component:
13
+ https://github.com/primeqa/primeqa/tree/main/examples/drdecr
14
 
15
  It is a Neural IR model built on top of the ColBERTv1 api and not directly compatible with Huggingface API. The inference result on XOR Dev dataset is:
16
  ```
17
+ R@2kt R@5kt
18
+ te 66.67 70.88
19
+ bn 70.23 75.08
20
+ fi 82.24 86.18
21
+ ja 65.92 72.93
22
+ ko 67.93 71.73
23
+ ru 63.07 69.71
24
+ ar 78.15 82.77
25
+ Avg 70.60 75.61
26
  ```
27
 
28
  # Limitations and Bias