nreimers commited on
Commit
e70081e
1 Parent(s): 6348f84

add captions dataset

Browse files
Files changed (2) hide show
  1. README.md +2 -1
  2. coco_captions.jsonl.gz +3 -0
README.md CHANGED
@@ -9,7 +9,7 @@ All files are in a `jsonl.gz` format: Each line contains a JSON-object that repr
9
  The JSON objects can come in different formats:
10
  - **Pairs:** `["text1", "text2"]` - This is a positive pair that should be close in vector space.
11
  - **Triplets:** `["anchor", "positive", "negative"]` - This is a triplet: The `positive` text should be close to the `anchor`, while the `negative` text should be distant to the `anchor`.
12
-
13
 
14
  ## Available Datasets
15
 
@@ -23,6 +23,7 @@ We measure the performance for each training dataset by training the [nreimers/M
23
  | --- | --- | :---: | :---: | --- |
24
  | [AllNLI.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/AllNLI.jsonl.gz) | Combination of SNLI + MultiNLI Triplets: (Anchor, Entailment_Text, Contradiction_Text) | 277,230 | 56.57 | [SNLI](https://huggingface.co/datasets/snli) and [MNLI](https://huggingface.co/datasets/multi_nli)
25
  | [altlex.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/altlex.jsonl.gz) | Matched pairs (English_Wikipedia, Simple_English_Wikipedia) | 112,696 | 55.95 | [altlex](https://github.com/chridey/altlex/)
 
26
  | [codesearchnet.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/codesearchnet.jsonl.gz) | CodeSearchNet corpus is a dataset of (comment, code) pairs from opensource libraries hosted on GitHub. It contains code and documentation for several programming languages. | 1,151,414 | 55.80 | [CodeSearchNet](https://huggingface.co/datasets/code_search_net)
27
  | [eli5_question_answer.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/eli5_question_answer.jsonl.gz) | (Question, Answer)-Pairs from ELI5 dataset | 325,475 | 58.24 | [ELI5](https://huggingface.co/datasets/eli5)
28
  | [fever_train.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/fever_train.jsonl.gz) | Training data from the FEVER corpus | 139,051 | 52.63 | [FEVER](https://huggingface.co/datasets/fever)
 
9
  The JSON objects can come in different formats:
10
  - **Pairs:** `["text1", "text2"]` - This is a positive pair that should be close in vector space.
11
  - **Triplets:** `["anchor", "positive", "negative"]` - This is a triplet: The `positive` text should be close to the `anchor`, while the `negative` text should be distant to the `anchor`.
12
+ - **Sets:** `{"set": ["text1", "text2", ...]}` A set of texts describing the same thing, e.g. different paraphrases of the same question, different captions for the same image. Any combination of the elements is considered as a positive pair.
13
 
14
  ## Available Datasets
15
 
 
23
  | --- | --- | :---: | :---: | --- |
24
  | [AllNLI.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/AllNLI.jsonl.gz) | Combination of SNLI + MultiNLI Triplets: (Anchor, Entailment_Text, Contradiction_Text) | 277,230 | 56.57 | [SNLI](https://huggingface.co/datasets/snli) and [MNLI](https://huggingface.co/datasets/multi_nli)
25
  | [altlex.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/altlex.jsonl.gz) | Matched pairs (English_Wikipedia, Simple_English_Wikipedia) | 112,696 | 55.95 | [altlex](https://github.com/chridey/altlex/)
26
+ | [coco_captions.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/coco_captions.jsonl.gz) | Different captions for the same image | 82,783 | 53.77 | [COCO](https://cocodataset.org/)
27
  | [codesearchnet.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/codesearchnet.jsonl.gz) | CodeSearchNet corpus is a dataset of (comment, code) pairs from opensource libraries hosted on GitHub. It contains code and documentation for several programming languages. | 1,151,414 | 55.80 | [CodeSearchNet](https://huggingface.co/datasets/code_search_net)
28
  | [eli5_question_answer.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/eli5_question_answer.jsonl.gz) | (Question, Answer)-Pairs from ELI5 dataset | 325,475 | 58.24 | [ELI5](https://huggingface.co/datasets/eli5)
29
  | [fever_train.jsonl.gz](https://huggingface.co/datasets/sentence-transformers/embedding-training-data/resolve/main/fever_train.jsonl.gz) | Training data from the FEVER corpus | 139,051 | 52.63 | [FEVER](https://huggingface.co/datasets/fever)
coco_captions.jsonl.gz ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bf0a7a50a7a43f4f010690bf3f7365ef0ce98afd0ea5747d04d00c8f3917a5f8
3
+ size 6316394