Skip to content

Commit 40124bb

Browse files
committed
update default cross-encoder model to Xenova/ms-marco-MiniLM-L-6-v2 across docs
1 parent f2dda99 commit 40124bb

4 files changed

Lines changed: 12 additions & 12 deletions

File tree

docs/reference.md

Lines changed: 3 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -79,13 +79,13 @@ SEARCH docs SIMILAR TO 'hello' LIMIT 5
7979
### Cross-encoder reranking (RERANK default)
8080

8181
```
82-
cross-encoder/ms-marco-MiniLM-L-6-v2
82+
Xenova/ms-marco-MiniLM-L-6-v2
8383
```
8484

8585
| Model | Notes |
8686
|---|---|
87-
| `cross-encoder/ms-marco-MiniLM-L-6-v2` | Default. Fast passage reranker |
88-
| `cross-encoder/ms-marco-MiniLM-L-12-v2` | Larger, higher quality |
87+
| `Xenova/ms-marco-MiniLM-L-6-v2` | Default. Fast passage reranker |
88+
| `Xenova/ms-marco-MiniLM-L-12-v2` | Larger, higher quality |
8989
| `BAAI/bge-reranker-base` | Strong general-purpose reranker |
9090
| `BAAI/bge-reranker-large` | Highest quality, slower |
9191

docs/search.md

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -370,15 +370,15 @@ SEARCH articles SIMILAR TO 'attention mechanism in transformers' LIMIT 10 USING
370370
Custom cross-encoder model:
371371
```sql
372372
SEARCH articles SIMILAR TO 'semantic search' LIMIT 5
373-
RERANK MODEL 'cross-encoder/ms-marco-MiniLM-L-6-v2'
373+
RERANK MODEL 'Xenova/ms-marco-MiniLM-L-6-v2'
374374
```
375375

376-
**Default cross-encoder model:** `cross-encoder/ms-marco-MiniLM-L-6-v2`
376+
**Default cross-encoder model:** `Xenova/ms-marco-MiniLM-L-6-v2`
377377

378378
| Model | Notes |
379379
|---|---|
380-
| `cross-encoder/ms-marco-MiniLM-L-6-v2` | Default. Fast and accurate for passage reranking |
381-
| `cross-encoder/ms-marco-MiniLM-L-12-v2` | Larger, higher quality, slower |
380+
| `Xenova/ms-marco-MiniLM-L-6-v2` | Default. Fast and accurate for passage reranking |
381+
| `Xenova/ms-marco-MiniLM-L-12-v2` | Larger, higher quality, slower |
382382
| `BAAI/bge-reranker-base` | BGE reranker, strong general-purpose performance |
383383
| `BAAI/bge-reranker-large` | Highest quality BGE reranker, slower |
384384

resources/Features.md

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -431,7 +431,7 @@ SEARCH medical_records SIMILAR TO 'neuromuscular junction antibody acetylcholine
431431
AND sub_specialty IN ('neuromuscular', 'neuroimmunology')
432432
AND severity IN ('high', 'critical')
433433
AND year >= 2022
434-
RERANK MODEL 'cross-encoder/ms-marco-MiniLM-L-6-v2'
434+
RERANK MODEL 'Xenova/ms-marco-MiniLM-L-6-v2'
435435
```
436436

437437
---

tests/test_parser.py

Lines changed: 4 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -1003,10 +1003,10 @@ def test_rerank_flag_set(self):
10031003

10041004
def test_rerank_with_model(self):
10051005
node = parse(
1006-
"SEARCH col SIMILAR TO 'q' LIMIT 5 RERANK MODEL 'cross-encoder/ms-marco-MiniLM-L-6-v2'"
1006+
"SEARCH col SIMILAR TO 'q' LIMIT 5 RERANK MODEL 'Xenova/ms-marco-MiniLM-L-6-v2'"
10071007
)
10081008
assert node.rerank is True
1009-
assert node.rerank_model == "cross-encoder/ms-marco-MiniLM-L-6-v2"
1009+
assert node.rerank_model == "Xenova/ms-marco-MiniLM-L-6-v2"
10101010

10111011
def test_rerank_default_false(self):
10121012
node = parse("SEARCH col SIMILAR TO 'q' LIMIT 5")
@@ -1032,12 +1032,12 @@ def test_rerank_with_where(self):
10321032
def test_rerank_with_hybrid_where_and_model(self):
10331033
node = parse(
10341034
"SEARCH col SIMILAR TO 'q' LIMIT 5 USING HYBRID WHERE year > 2020 "
1035-
"RERANK MODEL 'cross-encoder/ms-marco-MiniLM-L-6-v2'"
1035+
"RERANK MODEL 'Xenova/ms-marco-MiniLM-L-6-v2'"
10361036
)
10371037
assert node.hybrid is True
10381038
assert node.query_filter is not None
10391039
assert node.rerank is True
1040-
assert node.rerank_model == "cross-encoder/ms-marco-MiniLM-L-6-v2"
1040+
assert node.rerank_model == "Xenova/ms-marco-MiniLM-L-6-v2"
10411041

10421042
def test_rerank_lowercase(self):
10431043
node = parse("SEARCH col SIMILAR TO 'q' LIMIT 5 rerank")

0 commit comments

Comments
 (0)