Roman Solomatin
commited on
Commit
•
5936f84
1
Parent(s):
ef17c1b
fix FloresBitextMining split
Browse files- paths.json +46 -46
- results.py +4 -0
paths.json
CHANGED
@@ -8906,52 +8906,6 @@
|
|
8906 |
"results/bge-small-zh-v1.5/no_revision_available/CMedQAv2.json",
|
8907 |
"results/bge-small-zh-v1.5/no_revision_available/STSB.json"
|
8908 |
],
|
8909 |
-
"cointegrated__LaBSE-en-ru": [
|
8910 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuBQRetrieval.json",
|
8911 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MassiveIntentClassification.json",
|
8912 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/InappropriatenessClassification.json",
|
8913 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuReviewsClassification.json",
|
8914 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/SIB200ClusteringS2S.json",
|
8915 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringP2P.json",
|
8916 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/BelebeleRetrieval.json",
|
8917 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchGRNTIClusteringP2P.json",
|
8918 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/LanguageClassification.json",
|
8919 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringS2S.v2.json",
|
8920 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/TERRa.json",
|
8921 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/Tatoeba.json",
|
8922 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSTSBenchmarkSTS.json",
|
8923 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/CyrillicTurkicLangClassification.json",
|
8924 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RiaNewsRetrieval.json",
|
8925 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/GeoreviewClassification.json",
|
8926 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/SensitiveTopicsClassification.json",
|
8927 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/KinopoiskClassification.json",
|
8928 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RUParaPhraserSTS.json",
|
8929 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringS2S.json",
|
8930 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MassiveScenarioClassification.json",
|
8931 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchOECDClusteringP2P.json",
|
8932 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MultilingualSentimentClassification.json",
|
8933 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/PublicHealthQA.json",
|
8934 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/FloresBitextMining.json",
|
8935 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchGRNTIClassification.json",
|
8936 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/HeadlineClassification.json",
|
8937 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/XQuADRetrieval.json",
|
8938 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/BUCC.v2.json",
|
8939 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/XNLI.json",
|
8940 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/XNLIV2.json",
|
8941 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/STS22.json",
|
8942 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MultiLongDocRetrieval.json",
|
8943 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/CEDRClassification.json",
|
8944 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/SIB200Classification.json",
|
8945 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/GeoreviewClusteringP2P.json",
|
8946 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringP2P.v2.json",
|
8947 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/STS22.v2.json",
|
8948 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuBQReranking.json",
|
8949 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/NTREXBitextMining.json",
|
8950 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/OpusparcusPC.json",
|
8951 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchOECDClassification.json",
|
8952 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/BibleNLPBitextMining.json",
|
8953 |
-
"results/cointegrated__LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/STSBenchmarkMultilingualSTS.json"
|
8954 |
-
],
|
8955 |
"elser-v2": [
|
8956 |
"results/elser-v2/no_revision_available/MassiveIntentClassification.json",
|
8957 |
"results/elser-v2/no_revision_available/BiorxivClusteringP2P.json",
|
@@ -9028,6 +8982,52 @@
|
|
9028 |
"results/gelectra-large/no_revision_available/TenKGnadClusteringP2P.json",
|
9029 |
"results/gelectra-large/no_revision_available/BlurbsClusteringS2S.json"
|
9030 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
9031 |
"text-embedding-3-small": [
|
9032 |
"results/text-embedding-3-small/no_revision_available/AlphaNLI.json",
|
9033 |
"results/text-embedding-3-small/no_revision_available/WinoGrande.json",
|
|
|
8906 |
"results/bge-small-zh-v1.5/no_revision_available/CMedQAv2.json",
|
8907 |
"results/bge-small-zh-v1.5/no_revision_available/STSB.json"
|
8908 |
],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8909 |
"elser-v2": [
|
8910 |
"results/elser-v2/no_revision_available/MassiveIntentClassification.json",
|
8911 |
"results/elser-v2/no_revision_available/BiorxivClusteringP2P.json",
|
|
|
8982 |
"results/gelectra-large/no_revision_available/TenKGnadClusteringP2P.json",
|
8983 |
"results/gelectra-large/no_revision_available/BlurbsClusteringS2S.json"
|
8984 |
],
|
8985 |
+
"LaBSE-en-ru": [
|
8986 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuBQRetrieval.json",
|
8987 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MassiveIntentClassification.json",
|
8988 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/InappropriatenessClassification.json",
|
8989 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuReviewsClassification.json",
|
8990 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/SIB200ClusteringS2S.json",
|
8991 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringP2P.json",
|
8992 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/BelebeleRetrieval.json",
|
8993 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchGRNTIClusteringP2P.json",
|
8994 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/LanguageClassification.json",
|
8995 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringS2S.v2.json",
|
8996 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/TERRa.json",
|
8997 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/Tatoeba.json",
|
8998 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSTSBenchmarkSTS.json",
|
8999 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/CyrillicTurkicLangClassification.json",
|
9000 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RiaNewsRetrieval.json",
|
9001 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/GeoreviewClassification.json",
|
9002 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/SensitiveTopicsClassification.json",
|
9003 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/KinopoiskClassification.json",
|
9004 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RUParaPhraserSTS.json",
|
9005 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringS2S.json",
|
9006 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MassiveScenarioClassification.json",
|
9007 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchOECDClusteringP2P.json",
|
9008 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MultilingualSentimentClassification.json",
|
9009 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/PublicHealthQA.json",
|
9010 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/FloresBitextMining.json",
|
9011 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchGRNTIClassification.json",
|
9012 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/HeadlineClassification.json",
|
9013 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/XQuADRetrieval.json",
|
9014 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/BUCC.v2.json",
|
9015 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/XNLI.json",
|
9016 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/XNLIV2.json",
|
9017 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/STS22.json",
|
9018 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MultiLongDocRetrieval.json",
|
9019 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/CEDRClassification.json",
|
9020 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/SIB200Classification.json",
|
9021 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/GeoreviewClusteringP2P.json",
|
9022 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/MLSUMClusteringP2P.v2.json",
|
9023 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/STS22.v2.json",
|
9024 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuBQReranking.json",
|
9025 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/NTREXBitextMining.json",
|
9026 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/OpusparcusPC.json",
|
9027 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/RuSciBenchOECDClassification.json",
|
9028 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/BibleNLPBitextMining.json",
|
9029 |
+
"results/LaBSE-en-ru/cf0714e606d4af551e14ad69a7929cd6b0da7f7e/STSBenchmarkMultilingualSTS.json"
|
9030 |
+
],
|
9031 |
"text-embedding-3-small": [
|
9032 |
"results/text-embedding-3-small/no_revision_available/AlphaNLI.json",
|
9033 |
"results/text-embedding-3-small/no_revision_available/WinoGrande.json",
|
results.py
CHANGED
@@ -38,6 +38,8 @@ DEV_SPLIT = ["CmedqaRetrieval", "CovidRetrieval", "DuRetrieval", "EcomRetrieval"
|
|
38 |
TESTFULL_SPLIT = ["OpusparcusPC"]
|
39 |
# Use "standard" split
|
40 |
STANDARD_SPLIT = ["BrightRetrieval"]
|
|
|
|
|
41 |
|
42 |
TEST_AVG_SPLIT = {
|
43 |
"LEMBNeedleRetrieval": ["test_256", "test_512", "test_1024", "test_2048", "test_4096", "test_8192", "test_16384", "test_32768"],
|
@@ -333,6 +335,8 @@ class MTEBResults(datasets.GeneratorBasedBuilder):
|
|
333 |
split = "test.full"
|
334 |
elif (ds_name in STANDARD_SPLIT) and ("standard" in res_dict):
|
335 |
split = "standard"
|
|
|
|
|
336 |
elif (ds_name in TEST_AVG_SPLIT):
|
337 |
# Average splits
|
338 |
res_dict["test_avg"] = {}
|
|
|
38 |
TESTFULL_SPLIT = ["OpusparcusPC"]
|
39 |
# Use "standard" split
|
40 |
STANDARD_SPLIT = ["BrightRetrieval"]
|
41 |
+
# Use "devtest" split
|
42 |
+
DEVTEST_SPLIT = ["FloresBitextMining"]
|
43 |
|
44 |
TEST_AVG_SPLIT = {
|
45 |
"LEMBNeedleRetrieval": ["test_256", "test_512", "test_1024", "test_2048", "test_4096", "test_8192", "test_16384", "test_32768"],
|
|
|
335 |
split = "test.full"
|
336 |
elif (ds_name in STANDARD_SPLIT) and ("standard" in res_dict):
|
337 |
split = "standard"
|
338 |
+
elif (ds_name in DEVTEST_SPLIT) and ("standard" in res_dict):
|
339 |
+
split = "devtest"
|
340 |
elif (ds_name in TEST_AVG_SPLIT):
|
341 |
# Average splits
|
342 |
res_dict["test_avg"] = {}
|