Jo Kristian Bergum
commited on
Commit
•
d7e79ea
1
Parent(s):
18e6702
Add recall
Browse files
README.md
CHANGED
@@ -19,8 +19,18 @@ To use this model with vespa.ai for MS Marco Passage Ranking, see
|
|
19 |
|------------------------------------|----------------|
|
20 |
| Dev | 0.364 |
|
21 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
22 |
The MRR@10 on dev is achieved by re-ranking 1K retrieved by a dense retriever based on
|
23 |
-
[sentence-transformers/msmarco-MiniLM-L-6-v3](https://huggingface.co/sentence-transformers/msmarco-MiniLM-L-6-v3).
|
|
|
24 |
|
25 |
The official baseline BM25 ranking model MRR@10 0.16 on eval and 0.167 on dev question set.
|
26 |
See [MS Marco Passage Ranking Leaderboard](https://microsoft.github.io/msmarco/).
|
|
|
19 |
|------------------------------------|----------------|
|
20 |
| Dev | 0.364 |
|
21 |
|
22 |
+
Recall@k On Dev (6980 queries)
|
23 |
+
|
24 |
+
|K | Recall@K |
|
25 |
+
|------------------------------------|----------------|
|
26 |
+
| 50 | 0.816 |
|
27 |
+
| 200 | 0.905 |
|
28 |
+
| 1000 | 0.939 |
|
29 |
+
|
30 |
+
|
31 |
The MRR@10 on dev is achieved by re-ranking 1K retrieved by a dense retriever based on
|
32 |
+
[sentence-transformers/msmarco-MiniLM-L-6-v3](https://huggingface.co/sentence-transformers/msmarco-MiniLM-L-6-v3). Re-ranking the original top 1000 dev
|
33 |
+
is 0.354 MRR@10 (Recall@1K 0.82).
|
34 |
|
35 |
The official baseline BM25 ranking model MRR@10 0.16 on eval and 0.167 on dev question set.
|
36 |
See [MS Marco Passage Ranking Leaderboard](https://microsoft.github.io/msmarco/).
|