Hugging Face's logo

Datasets:
hotchpotch
/
JQaRA

Tasks:
Question Answering
Modalities:
Text
Formats:
parquet
Languages:
Japanese
Size:
100K - 1M
Libraries:
Datasets
pandas
License:
Dataset card Data Studio Files Community
1
hotchpotch commited on
Commit
7c4a09e
·
verified ·
1 Parent(s): eccb8d0

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +29 -27
README.md CHANGED
@@ -1,6 +1,7 @@
1
  ---
2
 
3
 
 
4
  dataset_info:
5
  features:
6
  - name: id
@@ -73,45 +74,46 @@ JQaRA は質問に対して、候補となる 100 件のデータ(一件以上
73
 
74
  ### 評価結果
75
 
76
- 以下はさまざまなモデルの評価結果です。評価は nDCG@10 以外にも、参考までにいくつかの指標を表示しています。
77
 
78
  #### 文ベクトルモデル
79
 
80
- | model_names | ndcg@10 | mrr@10 | ndcg@100 | mrr@100 |
81
- | :------------------------------------------------------------------------------ | ------: | -----: | -------: | ------: |
82
- | [bge-m3+dense](https://huggingface.co/BAAI/bge-m3) | 0.539 | 0.7854 | 0.7206 | 0.7884 |
83
- | [fio-base-japanese-v0.1](https://huggingface.co/bclavie/fio-base-japanese-v0.1) | 0.3718 | 0.6161 | 0.6077 | 0.6219 |
84
- | [sup-simcse-ja-base](https://huggingface.co/cl-nagoya/sup-simcse-ja-base) | 0.3237 | 0.5407 | 0.5722 | 0.5504 |
85
- | [sup-simcse-ja-large](https://huggingface.co/cl-nagoya/sup-simcse-ja-large) | 0.3571 | 0.575 | 0.5955 | 0.583 |
86
- | [unsup-simcse-ja-base](https://huggingface.co/cl-nagoya/unsup-simcse-ja-base) | 0.3121 | 0.5209 | 0.5664 | 0.5298 |
87
- | [unsup-simcse-ja-large](https://huggingface.co/cl-nagoya/unsup-simcse-ja-large) | 0.3928 | 0.6257 | 0.6217 | 0.6321 |
88
- | [multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) | 0.554 | 0.7988 | 0.731 | 0.8011 |
89
- | [multilingual-e5-small](https://huggingface.co/intfloat/multilingual-e5-small) | 0.4917 | 0.7291 | 0.6894 | 0.7325 |
90
- | [GLuCoSE-base-ja](https://huggingface.co/pkshatech/GLuCoSE-base-ja) | 0.3085 | 0.5179 | 0.5643 | 0.5268 |
91
- | [text-embedding-3-small](https://platform.openai.com/docs/guides/embeddings) | 0.3881 | 0.6107 | 0.6173 | 0.6171 |
92
 
93
  #### ColBERT モデル
94
 
95
- | model_names | ndcg@10 | mrr@10 | ndcg@100 | mrr@100 |
96
- | :---------------------------------------------------- | ------: | -----: | -------: | ------: |
97
- | [bge-m3+colbert](https://huggingface.co/BAAI/bge-m3) | 0.5656 | 0.8095 | 0.7371 | 0.8119 |
98
- | [JaColBERT](https://huggingface.co/bclavie/JaColBERT) | 0.5495 | 0.8108 | 0.7297 | 0.8137 |
 
99
 
100
  #### CrossEncoder モデル
101
 
102
- | model_names | ndcg@10 | mrr@10 | ndcg@100 | mrr@100 |
103
- | :---------------------------------------------------------------------------------------------------------------------- | ------: | -----: | -------: | ------: |
104
- | [bge-reranker-base](https://huggingface.co/BAAI/bge-reranker-base) | 0.2445 | 0.4378 | 0.522 | 0.4504 |
105
- | [bge-reranker-large](https://huggingface.co/BAAI/bge-reranker-large) | 0.4718 | 0.7108 | 0.6791 | 0.7149 |
106
- | [cross-encoder-mmarco-mMiniLMv2-L12-H384-v1](https://huggingface.co/corrius/cross-encoder-mmarco-mMiniLMv2-L12-H384-v1) | 0.5588 | 0.8107 | 0.7325 | 0.8139 |
107
 
108
  #### その他モデル
109
 
110
- | model_names | ndcg@10 | mrr@10 | ndcg@100 | mrr@100 |
111
- | :-------------------------------------------------- | ------: | -----: | -------: | ------: |
112
- | [bge-m3+all](https://huggingface.co/BAAI/bge-m3) | 0.576 | 0.8178 | 0.7454 | 0.8202 |
113
- | [bge-m3+sparse](https://huggingface.co/BAAI/bge-m3) | 0.5088 | 0.7596 | 0.7017 | 0.7626 |
114
- | bm25 | 0.458 | 0.702 | 0.6701 | 0.706 |
115
 
116
  ---
117
 
 
1
  ---
2
 
3
 
4
+
5
  dataset_info:
6
  features:
7
  - name: id
 
74
 
75
  ### 評価結果
76
 
77
+ 以下はさまざまなモデルの評価結果です。評価は nDCG@10 以外にも、参考まで MRR@10 の数値も掲載しています。
78
 
79
  #### 文ベクトルモデル
80
 
81
+ | model_names | ndcg@10 | mrr@10 |
82
+ | :------------------------------------------------------------------------------ | ------: | -----: |
83
+ | [bge-m3+dense](https://huggingface.co/BAAI/bge-m3) | 0.539 | 0.7854 |
84
+ | [fio-base-japanese-v0.1](https://huggingface.co/bclavie/fio-base-japanese-v0.1) | 0.3718 | 0.6161 |
85
+ | [sup-simcse-ja-base](https://huggingface.co/cl-nagoya/sup-simcse-ja-base) | 0.3237 | 0.5407 |
86
+ | [sup-simcse-ja-large](https://huggingface.co/cl-nagoya/sup-simcse-ja-large) | 0.3571 | 0.575 |
87
+ | [unsup-simcse-ja-base](https://huggingface.co/cl-nagoya/unsup-simcse-ja-base) | 0.3121 | 0.5209 |
88
+ | [unsup-simcse-ja-large](https://huggingface.co/cl-nagoya/unsup-simcse-ja-large) | 0.3928 | 0.6257 |
89
+ | [multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) | 0.554 | 0.7988 |
90
+ | [multilingual-e5-small](https://huggingface.co/intfloat/multilingual-e5-small) | 0.4917 | 0.7291 |
91
+ | [GLuCoSE-base-ja](https://huggingface.co/pkshatech/GLuCoSE-base-ja) | 0.3085 | 0.5179 |
92
+ | [text-embedding-3-small](https://platform.openai.com/docs/guides/embeddings) | 0.3881 | 0.6107 |
93
 
94
  #### ColBERT モデル
95
 
96
+ | model_names | ndcg@10 | mrr@10 |
97
+ | :-------------------------------------------------------- | ------: | -----: |
98
+ | [bge-m3+colbert](https://huggingface.co/BAAI/bge-m3) | 0.5656 | 0.8095 |
99
+ | [JaColBERT](https://huggingface.co/bclavie/JaColBERT) | 0.5495 | 0.8108 |
100
+ | [JaColBERTv2](https://huggingface.co/bclavie/JaColBERTv2) | 0.5847 | 0.8356 |
101
 
102
  #### CrossEncoder モデル
103
 
104
+ | model_names | ndcg@10 | mrr@10 |
105
+ | :---------------------------------------------------------------------------------------------------------------------- | ------: | -----: |
106
+ | [bge-reranker-base](https://huggingface.co/BAAI/bge-reranker-base) | 0.2445 | 0.4378 |
107
+ | [bge-reranker-large](https://huggingface.co/BAAI/bge-reranker-large) | 0.4718 | 0.7108 |
108
+ | [cross-encoder-mmarco-mMiniLMv2-L12-H384-v1](https://huggingface.co/corrius/cross-encoder-mmarco-mMiniLMv2-L12-H384-v1) | 0.5588 | 0.8107 |
109
 
110
  #### その他モデル
111
 
112
+ | model_names | ndcg@10 | mrr@10 |
113
+ | :-------------------------------------------------- | ------: | -----: |
114
+ | [bge-m3+all](https://huggingface.co/BAAI/bge-m3) | 0.576 | 0.8178 |
115
+ | [bge-m3+sparse](https://huggingface.co/BAAI/bge-m3) | 0.5088 | 0.7596 |
116
+ | bm25 | 0.458 | 0.702 |
117
 
118
  ---
119