-
Notifications
You must be signed in to change notification settings - Fork 290
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
fix: Added medical qa dataset (#333)
* Added news classification dataset. * Fixes on suggestions * Added new medical qa dataset * Update model run files and model path * Added points for dataset. * Fixes --------- Co-authored-by: Kenneth Enevoldsen <kennethcenevoldsen@gmail.com>
- Loading branch information
1 parent
84cffa2
commit 80acc3e
Showing
6 changed files
with
188 additions
and
9 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,42 @@ | ||
| from __future__ import annotations | ||
|
|
||
| from mteb.abstasks.AbsTaskRetrieval import AbsTaskRetrieval | ||
| from mteb.abstasks.TaskMetadata import TaskMetadata | ||
|
|
||
|
|
||
| class MedicalQARetrieval(AbsTaskRetrieval): | ||
| metadata = TaskMetadata( | ||
| name="MedicalQARetrieval", | ||
| description="The dataset consists 2048 medical question and answer pairs.", | ||
| reference="https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-019-3119-4", | ||
| dataset={ | ||
| "path": "mteb/medical_qa", | ||
| "revision": "ae763399273d8b20506b80cf6f6f9a31a6a2b238", | ||
| }, | ||
| type="Retrieval", | ||
| category="s2s", | ||
| eval_splits=["test"], | ||
| eval_langs=["en"], | ||
| main_score="ndcg_at_10", | ||
| date=("2017-01-01", "2019-12-31"), # best guess, | ||
| form=["written"], | ||
| domains=["Medical"], | ||
| task_subtypes=["Article retrieval"], | ||
| license="CC0 1.0 Universal", | ||
| socioeconomic_status="medium", | ||
| annotations_creators="derived", | ||
| dialect=[], | ||
| text_creation="found", | ||
| bibtex_citation="""@ARTICLE{BenAbacha-BMC-2019, | ||
| author = {Asma {Ben Abacha} and Dina Demner{-}Fushman}, | ||
| title = {A Question-Entailment Approach to Question Answering}, | ||
| journal = {{BMC} Bioinform.}, | ||
| volume = {20}, | ||
| number = {1}, | ||
| pages = {511:1--511:23}, | ||
| year = {2019}, | ||
| url = {https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-019-3119-4} | ||
| } """, | ||
| n_samples={"test": 2048}, | ||
| avg_character_length={"test": 1205.9619140625}, | ||
| ) |
38 changes: 38 additions & 0 deletions
38
results/intfloat__multilingual-e5-small/MedicalQARetrieval.json
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,38 @@ | ||
| { | ||
| "dataset_revision": "ae763399273d8b20506b80cf6f6f9a31a6a2b238", | ||
| "mteb_dataset_name": "MedicalQARetrieval", | ||
| "mteb_version": "1.5.4", | ||
| "test": { | ||
| "evaluation_time": 29.86, | ||
| "map_at_1": 0.43896, | ||
| "map_at_10": 0.56803, | ||
| "map_at_100": 0.57388, | ||
| "map_at_1000": 0.5741, | ||
| "map_at_3": 0.54167, | ||
| "map_at_5": 0.55658, | ||
| "mrr_at_1": 0.43896, | ||
| "mrr_at_10": 0.56815, | ||
| "mrr_at_100": 0.57401, | ||
| "mrr_at_1000": 0.57423, | ||
| "mrr_at_3": 0.54191, | ||
| "mrr_at_5": 0.55671, | ||
| "ndcg_at_1": 0.43896, | ||
| "ndcg_at_10": 0.62856, | ||
| "ndcg_at_100": 0.65666, | ||
| "ndcg_at_1000": 0.66229, | ||
| "ndcg_at_3": 0.57453, | ||
| "ndcg_at_5": 0.60145, | ||
| "precision_at_1": 0.43896, | ||
| "precision_at_10": 0.08174, | ||
| "precision_at_100": 0.00948, | ||
| "precision_at_1000": 0.00099, | ||
| "precision_at_3": 0.22314, | ||
| "precision_at_5": 0.14697, | ||
| "recall_at_1": 0.43896, | ||
| "recall_at_10": 0.81738, | ||
| "recall_at_100": 0.94824, | ||
| "recall_at_1000": 0.99219, | ||
| "recall_at_3": 0.66943, | ||
| "recall_at_5": 0.73486 | ||
| } | ||
| } |
38 changes: 38 additions & 0 deletions
38
results/sentence-transformers__paraphrase-multilingual-MiniLM-L12-v2/MedicalQARetrieval.json
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,38 @@ | ||
| { | ||
| "dataset_revision": "ae763399273d8b20506b80cf6f6f9a31a6a2b238", | ||
| "mteb_dataset_name": "MedicalQARetrieval", | ||
| "mteb_version": "1.5.4", | ||
| "test": { | ||
| "evaluation_time": 11.98, | ||
| "map_at_1": 0.35059, | ||
| "map_at_10": 0.47558, | ||
| "map_at_100": 0.48095, | ||
| "map_at_1000": 0.48147, | ||
| "map_at_3": 0.44621, | ||
| "map_at_5": 0.46586, | ||
| "mrr_at_1": 0.35059, | ||
| "mrr_at_10": 0.47562, | ||
| "mrr_at_100": 0.48099, | ||
| "mrr_at_1000": 0.4815, | ||
| "mrr_at_3": 0.44637, | ||
| "mrr_at_5": 0.4659, | ||
| "ndcg_at_1": 0.35059, | ||
| "ndcg_at_10": 0.5363, | ||
| "ndcg_at_100": 0.5641, | ||
| "ndcg_at_1000": 0.57894, | ||
| "ndcg_at_3": 0.47757, | ||
| "ndcg_at_5": 0.51276, | ||
| "precision_at_1": 0.35059, | ||
| "precision_at_10": 0.07261, | ||
| "precision_at_100": 0.0086, | ||
| "precision_at_1000": 0.00098, | ||
| "precision_at_3": 0.18945, | ||
| "precision_at_5": 0.13066, | ||
| "recall_at_1": 0.35059, | ||
| "recall_at_10": 0.72607, | ||
| "recall_at_100": 0.86035, | ||
| "recall_at_1000": 0.97949, | ||
| "recall_at_3": 0.56836, | ||
| "recall_at_5": 0.65332 | ||
| } | ||
| } |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,60 @@ | ||
| from __future__ import annotations | ||
|
|
||
| import os | ||
| import uuid | ||
| from typing import Dict | ||
|
|
||
| from datasets import load_dataset | ||
| from huggingface_hub import create_repo, upload_file | ||
|
|
||
|
|
||
| def preprocess_data(example: Dict) -> Dict: | ||
| """ | ||
| Preprocessed the data in a format easier | ||
| to handle for the loading of queries and corpus | ||
| ------ | ||
| PARAMS | ||
| example : element in med-qa dataset | ||
| """ | ||
| return { | ||
| "query-id": str(uuid.uuid4()), | ||
| "query_text": example["Question"], | ||
| "corpus-id": str(uuid.uuid4()), | ||
| "answer_text": example["Answer"], | ||
| } | ||
|
|
||
|
|
||
| repo_name = "mteb/medical_qa" | ||
| create_repo(repo_name, repo_type="dataset", token="") | ||
|
|
||
|
|
||
| raw_dset = load_dataset("keivalya/MedQuad-MedicalQnADataset") | ||
| dset = raw_dset["train"] | ||
| trimmed_dataset = dset.select(range(2048)) | ||
| updated_dataset = trimmed_dataset.map( | ||
| preprocess_data, remove_columns=["Question", "Answer", "qtype"] | ||
| ) | ||
| corpus_ds = updated_dataset.map( | ||
| lambda example: {"_id": example["corpus-id"], "text": example["answer_text"]}, | ||
| remove_columns=["query-id", "query_text", "corpus-id", "answer_text"], | ||
| ) | ||
| corpus_ds = corpus_ds.add_column("title", len(corpus_ds) * [""]) | ||
| default_ds = updated_dataset.map( | ||
| lambda example: example, remove_columns=["answer_text", "query_text"] | ||
| ) | ||
| default_ds = default_ds.add_column("score", len(corpus_ds) * [0]) | ||
| queries_ds = updated_dataset.map( | ||
| lambda example: {"_id": example["query-id"], "text": example["query_text"]}, | ||
| remove_columns=["corpus-id", "answer_text", "query-id", "query_text"], | ||
| ) | ||
| data = {"corpus": corpus_ds, "default": default_ds, "queries": queries_ds} | ||
| for splits in ["default", "queries"]: | ||
| save_path = f"{splits}.jsonl" | ||
| data[splits].to_json(save_path) | ||
| upload_file( | ||
| path_or_fileobj=save_path, | ||
| path_in_repo=save_path, | ||
| repo_id=repo_name, | ||
| repo_type="dataset", | ||
| ) | ||
| os.system(f"rm {save_path}") |