Rename

results/bge-base-zh/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.28242481419180376,
     "mrr": 0.2667063492063492
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/bge-large-zh-noinstruct/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.27102013701446426,
     "mrr": 0.26324206349206347
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/bge-large-zh/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.26232690178149815,
     "mrr": 0.25287301587301586
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/bge-small-zh/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.2282096733248694,
     "mrr": 0.21677380952380954
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/luotuo-bert-medium/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.1454654108166253,
     "mrr": 0.12848412698412698
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/m3e-base/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.17509891535107402,
     "mrr": 0.16049603174603175
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/m3e-large/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.16458504717332473,
     "mrr": 0.14978968253968256
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }

results/multilingual-e5-base/AFQMC.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "AFQMC",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": NaN,
+      "spearman": NaN
+    },
+    "euclidean": {
+      "pearson": NaN,
+      "spearman": NaN
+    },
+    "evaluation_time": 4.59,
+    "manhattan": {
+      "pearson": NaN,
+      "spearman": NaN
+    }
+  }
+}

results/multilingual-e5-base/ATEC.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "ATEC",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": 0.373550586257545,
+      "spearman": 0.3701257946883062
+    },
+    "euclidean": {
+      "pearson": 0.39803864674686384,
+      "spearman": 0.3696591560768722
+    },
+    "evaluation_time": 23.06,
+    "manhattan": {
+      "pearson": 0.39785415725026396,
+      "spearman": 0.3698413102943949
+    }
+  }
+}

results/multilingual-e5-base/BQ.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "BQ",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": 0.45303599781483117,
+      "spearman": 0.45448247854254875
+    },
+    "euclidean": {
+      "pearson": 0.4522686095980617,
+      "spearman": 0.4542176825905819
+    },
+    "evaluation_time": 12.41,
+    "manhattan": {
+      "pearson": 0.4532868838598347,
+      "spearman": 0.4551567659021363
+    }
+  }
+}

results/multilingual-e5-base/CLSClusteringP2P.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "CLSClusteringP2P",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "evaluation_time": 994.39,
+    "v_measure": 0.3240836272693557,
+    "v_measure_std": 0.014062578366647911
+  }
+}

results/multilingual-e5-base/CLSClusteringS2S.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "CLSClusteringS2S",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "evaluation_time": 68.17,
+    "v_measure": 0.36989955642239736,
+    "v_measure_std": 0.014148037653818231
+  }
+}

results/multilingual-e5-base/CMedQAv1.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "CMedQAv1",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "evaluation_time": 359.86,
+    "map": 0.6520611444006958,
+    "mrr": 0.7082547619047619
+  }
+}

results/multilingual-e5-base/CMedQAv2.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "CMedQAv2",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "evaluation_time": 345.36,
+    "map": 0.660602348976891,
+    "mrr": 0.7095924603174604
+  }
+}

results/multilingual-e5-base/Cmnli.json

@@ -0,0 +1,49 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "Cmnli",
+  "mteb_version": "1.1.1.dev0",
+  "validation": {
+    "cos_sim": {
+      "accuracy": 0.67083583884546,
+      "accuracy_threshold": 0.9160473346710205,
+      "ap": 0.7451262220072653,
+      "f1": 0.7099406528189912,
+      "f1_threshold": 0.8881622552871704,
+      "precision": 0.5882895343476257,
+      "recall": 0.895019873743278
+    },
+    "dot": {
+      "accuracy": 0.5526157546602526,
+      "accuracy_threshold": 260.6299743652344,
+      "ap": 0.5833343142596583,
+      "f1": 0.6792153125248194,
+      "f1_threshold": 170.95211791992188,
+      "precision": 0.5143132066394034,
+      "recall": 0.999766191255553
+    },
+    "euclidean": {
+      "accuracy": 0.6778111846061335,
+      "accuracy_threshold": 7.05479097366333,
+      "ap": 0.7498396139781417,
+      "f1": 0.7167630057803467,
+      "f1_threshold": 7.7890119552612305,
+      "precision": 0.6095362936260855,
+      "recall": 0.8697685293429974
+    },
+    "evaluation_time": 13.62,
+    "manhattan": {
+      "accuracy": 0.6754058929645219,
+      "accuracy_threshold": 155.12338256835938,
+      "ap": 0.7486919271065671,
+      "f1": 0.7158293838862558,
+      "f1_threshold": 173.24166870117188,
+      "precision": 0.6019448429778416,
+      "recall": 0.8828618190320318
+    },
+    "max": {
+      "accuracy": 0.6778111846061335,
+      "ap": 0.7498396139781417,
+      "f1": 0.7167630057803467
+    }
+  }
+}

results/multilingual-e5-base/IFlyTek.json

@@ -0,0 +1,13 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "IFlyTek",
+  "mteb_version": "1.1.1.dev0",
+  "validation": {
+    "accuracy": 0.4493266641015775,
+    "accuracy_stderr": 0.003512954893578556,
+    "evaluation_time": 254.82,
+    "f1": 0.36853572831286635,
+    "f1_stderr": 0.004218344207296586,
+    "main_score": 0.4493266641015775
+  }
+}

results/multilingual-e5-base/JDReview.json

@@ -0,0 +1,15 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "JDReview",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "accuracy": 0.7621013133208254,
+    "accuracy_stderr": 0.020668576891437772,
+    "ap": 0.3800848445284182,
+    "ap_stderr": 0.02019610804976287,
+    "evaluation_time": 15.17,
+    "f1": 0.6972380312259057,
+    "f1_stderr": 0.01793508902745797,
+    "main_score": 0.7621013133208254
+  }
+}

results/multilingual-e5-base/LCQMC.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "LCQMC",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": 0.6814095532091424,
+      "spearman": 0.741480433718059
+    },
+    "euclidean": {
+      "pearson": 0.7294175071690869,
+      "spearman": 0.7422597401119065
+    },
+    "evaluation_time": 9.45,
+    "manhattan": {
+      "pearson": 0.7288460318935632,
+      "spearman": 0.7414605912619334
+    }
+  }
+}

results/multilingual-e5-base/MMarcoReranking.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "dev": {
+    "evaluation_time": 367.86,
+    "map": 0.21758599602003656,
+    "mrr": 0.20473412698412694
+  },
+  "mteb_dataset_name": "MMarcoReranking",
+  "mteb_version": "1.1.1.dev0"
+}

results/multilingual-e5-base/MultilingualSentiment.json

@@ -0,0 +1,13 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "MultilingualSentiment",
+  "mteb_version": "1.1.1.dev0",
+  "validation": {
+    "accuracy": 0.6528,
+    "accuracy_stderr": 0.015590737991227706,
+    "evaluation_time": 23.58,
+    "f1": 0.6522681687065506,
+    "f1_stderr": 0.016840153401201912,
+    "main_score": 0.6528
+  }
+}

results/multilingual-e5-base/Ocnli.json

@@ -0,0 +1,49 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "Ocnli",
+  "mteb_version": "1.1.1.dev0",
+  "validation": {
+    "cos_sim": {
+      "accuracy": 0.5804006497022198,
+      "accuracy_threshold": 0.9106633067131042,
+      "ap": 0.5962534282768569,
+      "f1": 0.6782231852654387,
+      "f1_threshold": 0.8347713947296143,
+      "precision": 0.5153677277716795,
+      "recall": 0.9915522703273495
+    },
+    "dot": {
+      "accuracy": 0.5154304277206281,
+      "accuracy_threshold": 220.6337127685547,
+      "ap": 0.5164804174436244,
+      "f1": 0.6783667621776504,
+      "f1_threshold": 208.66842651367188,
+      "precision": 0.5132791327913279,
+      "recall": 1.0
+    },
+    "euclidean": {
+      "accuracy": 0.5885219274499188,
+      "accuracy_threshold": 7.201763153076172,
+      "ap": 0.6039135198006501,
+      "f1": 0.6795058139534884,
+      "f1_threshold": 9.642261505126953,
+      "precision": 0.518005540166205,
+      "recall": 0.9873284054910243
+    },
+    "evaluation_time": 2.96,
+    "manhattan": {
+      "accuracy": 0.5912290200324851,
+      "accuracy_threshold": 158.3519287109375,
+      "ap": 0.6046932716509501,
+      "f1": 0.6792316056542226,
+      "f1_threshold": 214.6934814453125,
+      "precision": 0.5171081677704195,
+      "recall": 0.989440337909187
+    },
+    "max": {
+      "accuracy": 0.5912290200324851,
+      "ap": 0.6046932716509501,
+      "f1": 0.6795058139534884
+    }
+  }
+}

results/multilingual-e5-base/OnlineShopping.json

@@ -0,0 +1,15 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "OnlineShopping",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "accuracy": 0.884,
+    "accuracy_stderr": 0.021004761364985816,
+    "ap": 0.8588925493540545,
+    "ap_stderr": 0.023481667516592298,
+    "evaluation_time": 13.69,
+    "f1": 0.8839123392161612,
+    "f1_stderr": 0.020984685258322238,
+    "main_score": 0.884
+  }
+}

results/multilingual-e5-base/PAWSX.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "PAWSX",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": 0.12180372023646169,
+      "spearman": 0.12139273779665764
+    },
+    "euclidean": {
+      "pearson": 0.13834888637912623,
+      "spearman": 0.1135323752139316
+    },
+    "evaluation_time": 4.54,
+    "manhattan": {
+      "pearson": 0.13870252363945929,
+      "spearman": 0.11384069303176406
+    }
+  }
+}

results/multilingual-e5-base/QBQTC.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "QBQTC",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": 0.26523520497461966,
+      "spearman": 0.2881374545462317
+    },
+    "euclidean": {
+      "pearson": 0.27024025636798843,
+      "spearman": 0.28985097167369306
+    },
+    "evaluation_time": 7.8,
+    "manhattan": {
+      "pearson": 0.2720676214772048,
+      "spearman": 0.29194526688581
+    }
+  }
+}

results/multilingual-e5-base/STS22.json

@@ -0,0 +1,22 @@
+{
+  "dataset_revision": "6d1ba47164174a496b7fa5d3569dae26a6813b80",
+  "mteb_dataset_name": "STS22",
+  "mteb_version": "1.1.0",
+  "test": {
+    "evaluation_time": 15.78,
+    "zh": {
+      "cos_sim": {
+        "pearson": 0.580593994248265,
+        "spearman": 0.6564042377559131
+      },
+      "euclidean": {
+        "pearson": 0.4745454874584024,
+        "spearman": 0.6263254821825651
+      },
+      "manhattan": {
+        "pearson": 0.47547561695110807,
+        "spearman": 0.6245144767699952
+      }
+    }
+  }
+}

results/multilingual-e5-base/STSB.json

@@ -0,0 +1,20 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "STSB",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "cos_sim": {
+      "pearson": 0.7923536624786671,
+      "spearman": 0.7904542774401909
+    },
+    "euclidean": {
+      "pearson": 0.7848397831159944,
+      "spearman": 0.7878225614423001
+    },
+    "evaluation_time": 2.11,
+    "manhattan": {
+      "pearson": 0.7831002065895636,
+      "spearman": 0.7857187450963864
+    }
+  }
+}

results/multilingual-e5-base/T2Reranking.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "dev": {
+    "evaluation_time": 1139.17,
+    "map": 0.643866753565647,
+    "mrr": 0.725825421005685
+  },
+  "mteb_dataset_name": "T2Reranking",
+  "mteb_version": "1.1.1.dev0"
+}

results/multilingual-e5-base/TNews.json

@@ -0,0 +1,13 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "TNews",
+  "mteb_version": "1.1.1.dev0",
+  "validation": {
+    "accuracy": 0.47056000000000003,
+    "accuracy_stderr": 0.008537470351339442,
+    "evaluation_time": 30.17,
+    "f1": 0.45538348928071193,
+    "f1_stderr": 0.008124697989629562,
+    "main_score": 0.47056000000000003
+  }
+}

results/multilingual-e5-base/ThuNewsClusteringP2P.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "ThuNewsClusteringP2P",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "evaluation_time": 445.69,
+    "v_measure": 0.40978660903187053,
+    "v_measure_std": 0.020457578952846982
+  }
+}

results/multilingual-e5-base/ThuNewsClusteringS2S.json

@@ -0,0 +1,10 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "ThuNewsClusteringS2S",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "evaluation_time": 59.43,
+    "v_measure": 0.5236140107043674,
+    "v_measure_std": 0.020888120540309855
+  }
+}

results/multilingual-e5-base/Waimai.json

@@ -0,0 +1,15 @@
+{
+  "dataset_revision": null,
+  "mteb_dataset_name": "Waimai",
+  "mteb_version": "1.1.1.dev0",
+  "test": {
+    "accuracy": 0.8442000000000001,
+    "accuracy_stderr": 0.013075167302944933,
+    "ap": 0.6580622339164859,
+    "ap_stderr": 0.021630020878655898,
+    "evaluation_time": 6.05,
+    "f1": 0.8243508593454243,
+    "f1_stderr": 0.012334290325224843,
+    "main_score": 0.8442000000000001
+  }
+}

results/text-embedding-ada-002/merge_cqadupstack.py

@@ -1,66 +0,0 @@
-"""
-Merges CQADupstack subset results
-Usage: python merge_cqadupstack.py path_to_results_folder
-"""
-import glob
-import json
-import logging
-import os
-import sys
-
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger(__name__)
-
-TASK_LIST_CQA = [
-    "CQADupstackAndroidRetrieval",
-    "CQADupstackEnglishRetrieval",
-    "CQADupstackGamingRetrieval",
-    "CQADupstackGisRetrieval",
-    "CQADupstackMathematicaRetrieval",
-    "CQADupstackPhysicsRetrieval",
-    "CQADupstackProgrammersRetrieval",
-    "CQADupstackStatsRetrieval",
-    "CQADupstackTexRetrieval",
-    "CQADupstackUnixRetrieval",
-    "CQADupstackWebmastersRetrieval",
-    "CQADupstackWordpressRetrieval",
-]
-
-NOAVG_KEYS = [
-    "evaluation_time",
-    "mteb_version",
-    "mteb_dataset_name",
-    "dataset_revision",
-]
-
-
-results_folder = sys.argv[1]
-# Ensure at least 1 character btw CQADupstack & Retrieval
-files = glob.glob(f'{results_folder.strip("/")}/CQADupstack*?*Retrieval.json')
-
-logger.info(f"Found CQADupstack files: {files}")
-
-if len(files) == len(TASK_LIST_CQA):
-    all_results = {}
-    for file_name in files:
-        with open(file_name, 'r', encoding='utf-8') as f:
-            results = json.load(f)
-            for split, split_results in results.items():
-                if split not in ("train", "validation", "dev", "test"):
-                    all_results[split] = split_results
-                    continue
-                all_results.setdefault(split, {})
-                for metric, score in split_results.items():
-                    all_results[split].setdefault(metric, 0)
-                    if metric == "evaluation_time":
-                        score = all_results[split][metric] + score
-                    elif metric not in NOAVG_KEYS:
-                        score = all_results[split][metric] + score * 1/len(TASK_LIST_CQA)
-                    all_results[split][metric] = score
-    all_results["mteb_dataset_name"] = "CQADupstackRetrieval"
-
-    logger.info("Saving ", all_results)
-    with open(os.path.join(results_folder, "CQADupstackRetrieval.json"), 'w', encoding='utf-8') as f:
-        json.dump(all_results, f, indent=4)
-else:
-    logger.warning(f"Got {len(files)}, but expected {len(TASK_LIST_CQA)} files. Missing: {set(TASK_LIST_CQA) - set([x.split('/')[-1].split('.')[0] for x in files])}; Too much: {set([x.split('/')[-1].split('.')[0] for x in files]) - set(TASK_LIST_CQA)}")

results/text2vec-large-chinese/MmarcoReranking.json

@@ -5,6 +5,6 @@
     "map": 0.12481971498981873,
     "mrr": 0.11227777777777778
   },
-  "mteb_dataset_name": "MmarcoReranking",
+  "mteb_dataset_name": "MMarcoReranking",
   "mteb_version": "1.0.2"
 }