xiaowenbin
commited on
Commit
•
dc25e36
1
Parent(s):
d92aead
Upload mteb_eval.py
Browse files- mteb_eval.py +18 -0
mteb_eval.py
ADDED
@@ -0,0 +1,18 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import logging
|
2 |
+
import functools
|
3 |
+
from mteb import MTEB
|
4 |
+
from sentence_transformers import SentenceTransformer
|
5 |
+
logging.basicConfig(level=logging.INFO)
|
6 |
+
logger = logging.getLogger("main")
|
7 |
+
|
8 |
+
# task_list
|
9 |
+
task_list = ['Classification', 'Clustering', 'Reranking', 'Retrieval', 'STS', 'PairClassification']
|
10 |
+
# languages
|
11 |
+
task_langs=["zh", "zh-CN"]
|
12 |
+
|
13 |
+
model_name = "DMetaSoul/Dmeta-embedding"
|
14 |
+
model = SentenceTransformer(model_name)
|
15 |
+
# normalize_embeddings should be true for this model
|
16 |
+
model.encode = functools.partial(model.encode, normalize_embeddings=True)
|
17 |
+
evaluation = MTEB(task_types=task_list, task_langs=task_langs)
|
18 |
+
evaluation.run(model, output_folder=f"results/zh/{model_name.split('/')[-1]}")
|