Update README.md
Browse files
README.md
CHANGED
|
@@ -59,7 +59,30 @@ This structure matches [LongEmbed benchmark](https://github.com/dwzhu-pku/LongEm
|
|
| 59 |
|
| 60 |
You can quickly try out document-level retrieval using BM25 with the following code snippet:
|
| 61 |
|
| 62 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 63 |
|
| 64 |
## License
|
| 65 |
The dataset is distributed under the Creative Commons Attribution-ShareAlike 3.0 License (CC BY-SA 3.0). You must give appropriate credit and share any derivative works under the same terms.
|
|
|
|
| 59 |
|
| 60 |
You can quickly try out document-level retrieval using BM25 with the following code snippet:
|
| 61 |
|
| 62 |
+
https://gist.github.com/ihainan/a1cf382c6042b90c8e55fe415f1b29e8
|
| 63 |
+
|
| 64 |
+
Usage:
|
| 65 |
+
|
| 66 |
+
```
|
| 67 |
+
$ python test_long_embed_bm25.py /home/ihainan/projects/Large/AI/DRCD-Simplified-Chinese/ir_dataset/train
|
| 68 |
+
...0
|
| 69 |
+
Building prefix dict from the default dictionary ...
|
| 70 |
+
Loading model from cache /tmp/jieba.cache
|
| 71 |
+
Loading model cost 0.404 seconds.
|
| 72 |
+
Prefix dict has been built successfully.
|
| 73 |
+
...200
|
| 74 |
+
...400
|
| 75 |
+
...600
|
| 76 |
+
...800
|
| 77 |
+
...1000
|
| 78 |
+
...1200
|
| 79 |
+
...1400
|
| 80 |
+
...1600
|
| 81 |
+
...1800
|
| 82 |
+
|
| 83 |
+
Acc@1: 64.76%
|
| 84 |
+
nDCG@10: 76.61%
|
| 85 |
+
```
|
| 86 |
|
| 87 |
## License
|
| 88 |
The dataset is distributed under the Creative Commons Attribution-ShareAlike 3.0 License (CC BY-SA 3.0). You must give appropriate credit and share any derivative works under the same terms.
|