rust/lancedb
1
0
Fork 0
mirror of https://github.com/lancedb/lancedb.git synced 2026-08-18 20:18:37 +00:00
Code Issues Packages Projects Releases Wiki Activity

chore(python): update Colbert architecture and minor improvements (#1547)

Browse Source 
- Update ColBertReranker architecture: The current implementation
doesn't use the right arch. This PR uses the implementation in Rerankers
library. Fixes https://github.com/lancedb/lancedb/issues/1546
Benchmark diff (hit rate):
Hybrid - 91 vs 87
reranked vector - 85 vs 80

- Reranking in FTS is basically disabled in main after last week's FTS
updates. I think there's no blocker in supporting that?
- Allow overriding accelerators: Most transformer based Rerankers and
Embedding automatically select device. This PR allows overriding those
settings by passing `device`. Fixes:
https://github.com/lancedb/lancedb/issues/1487

---------

Co-authored-by: BubbleCal <bubble-cal@outlook.com>
This commit is contained in:
Ayush Chaurasia
2024-08-21 12:26:52 +05:30
committed by GitHub
parent 85bb7e54e4
commit 7d65dd97cf
6 changed files with 37 additions and 69 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/python/tests/test_rerankers.py
+15 -10
View File
@@ -236,33 +236,37 @@ def test_rrf_reranker(tmp_path, use_tantivy):
@pytest.mark.skipif(
os.environ.get("COHERE_API_KEY") is None, reason="COHERE_API_KEY not set"
)
def test_cohere_reranker(tmp_path):
@pytest.mark.parametrize("use_tantivy", [True, False])
def test_cohere_reranker(tmp_path, use_tantivy):
pytest.importorskip("cohere")
reranker = CohereReranker()
table, schema = get_test_table(tmp_path)
table, schema = get_test_table(tmp_path, use_tantivy)
_run_test_reranker(reranker, table, "single player experience", None, schema)
def test_cross_encoder_reranker(tmp_path):
@pytest.mark.parametrize("use_tantivy", [True, False])
def test_cross_encoder_reranker(tmp_path, use_tantivy):
pytest.importorskip("sentence_transformers")
reranker = CrossEncoderReranker()
table, schema = get_test_table(tmp_path)
table, schema = get_test_table(tmp_path, use_tantivy)
_run_test_reranker(reranker, table, "single player experience", None, schema)
def test_colbert_reranker(tmp_path):
@pytest.mark.parametrize("use_tantivy", [True, False])
def test_colbert_reranker(tmp_path, use_tantivy):
pytest.importorskip("transformers")
reranker = ColbertReranker()
table, schema = get_test_table(tmp_path)
table, schema = get_test_table(tmp_path, use_tantivy)
_run_test_reranker(reranker, table, "single player experience", None, schema)
@pytest.mark.skipif(
os.environ.get("OPENAI_API_KEY") is None, reason="OPENAI_API_KEY not set"
)
def test_openai_reranker(tmp_path):
@pytest.mark.parametrize("use_tantivy", [True, False])
def test_openai_reranker(tmp_path, use_tantivy):
pytest.importorskip("openai")
table, schema = get_test_table(tmp_path)
table, schema = get_test_table(tmp_path, use_tantivy)
reranker = OpenaiReranker()
_run_test_reranker(reranker, table, "single player experience", None, schema)
@@ -270,8 +274,9 @@ def test_openai_reranker(tmp_path):
@pytest.mark.skipif(
os.environ.get("JINA_API_KEY") is None, reason="JINA_API_KEY not set"
)
def test_jina_reranker(tmp_path):
@pytest.mark.parametrize("use_tantivy", [True, False])
def test_jina_reranker(tmp_path, use_tantivy):
pytest.importorskip("jina")
table, schema = get_test_table(tmp_path)
table, schema = get_test_table(tmp_path, use_tantivy)
reranker = JinaReranker()
_run_test_reranker(reranker, table, "single player experience", None, schema)