-
Notifications
You must be signed in to change notification settings - Fork 252
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
411e313
commit 74a2400
Showing
4 changed files
with
289 additions
and
0 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
50 changes: 50 additions & 0 deletions
50
mteb/tasks/Image/Any2AnyRetrieval/eng/VQA2IT2TRetrieval.py
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 +1,50 @@ | ||
from __future__ import annotations | ||
|
||
from mteb.abstasks.Image.AbsTaskAny2AnyRetrieval import AbsTaskAny2AnyRetrieval | ||
from mteb.abstasks.TaskMetadata import TaskMetadata | ||
|
||
|
||
class VQA2IT2TRetrieval(AbsTaskAny2AnyRetrieval): | ||
metadata = TaskMetadata( | ||
name="VQA2IT2TRetrieval", | ||
description="Retrieve the correct answer for a question about an image.", | ||
reference="https://openaccess.thecvf.com/content_cvpr_2017/html/Goyal_Making_the_v_CVPR_2017_paper.html", | ||
dataset={ | ||
"path": "JamieSJS/vqa-2", | ||
"revision": "addc919d80589a767212ca455657f4e6c55e71f8", | ||
"trust_remote_code": True, | ||
}, | ||
type="Retrieval", | ||
category="it2t", | ||
eval_splits=["test"], | ||
eval_langs=["eng-Latn"], | ||
main_score="ndcg_at_10", | ||
date=("2017-07-01", "2017-07-01"), | ||
domains=["Web"], | ||
task_subtypes=["Image Text Retrieval"], | ||
license="CC BY-SA 4.0", | ||
annotations_creators="derived", | ||
dialect=[], | ||
modalities=["text", "image"], | ||
sample_creation="found", | ||
bibtex_citation="""@InProceedings{Goyal_2017_CVPR, | ||
author = {Goyal, Yash and Khot, Tejas and Summers-Stay, Douglas and Batra, Dhruv and Parikh, Devi}, | ||
title = {Making the v in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answering}, | ||
booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, | ||
month = {July}, | ||
year = {2017} | ||
} | ||
""", | ||
descriptive_stats={ | ||
"n_samples": {"test": 214354}, | ||
"avg_character_length": { | ||
"test": { | ||
"average_document_length": 0.0, | ||
"average_query_length": 0.0, | ||
"num_documents": 2143540, | ||
"num_queries": 214354, | ||
"average_relevant_docs_per_query": 10, | ||
} | ||
}, | ||
}, | ||
) |
51 changes: 51 additions & 0 deletions
51
mteb/tasks/Image/Any2AnyRetrieval/eng/VizWizIT2TRetrieval.py
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 +1,51 @@ | ||
from __future__ import annotations | ||
|
||
from mteb.abstasks.Image.AbsTaskAny2AnyRetrieval import AbsTaskAny2AnyRetrieval | ||
from mteb.abstasks.TaskMetadata import TaskMetadata | ||
|
||
|
||
class VizWizIT2TRetrieval(AbsTaskAny2AnyRetrieval): | ||
metadata = TaskMetadata( | ||
name="VizWizIT2TRetrieval", | ||
description="Retrieve the correct answer for a question about an image.", | ||
reference="https://openaccess.thecvf.com/content_cvpr_2018/papers/Gurari_VizWiz_Grand_Challenge_CVPR_2018_paper.pdf", | ||
dataset={ | ||
"path": "JamieSJS/vizwiz", | ||
"revision": "723c6ffed034eb9dcbb44777e1a9ddf97c5ee28a", | ||
"trust_remote_code": True, | ||
}, | ||
type="Retrieval", | ||
category="it2t", | ||
eval_splits=["test"], | ||
eval_langs=["eng-Latn"], | ||
main_score="ndcg_at_10", | ||
date=("2018-01-01", "2018-01-01"), | ||
domains=["Web"], | ||
task_subtypes=["Image Text Retrieval"], | ||
license="CC BY-SA 4.0", | ||
annotations_creators="derived", | ||
dialect=[], | ||
modalities=["text", "image"], | ||
sample_creation="found", | ||
bibtex_citation="""@inproceedings{gurari2018vizwiz, | ||
title={Vizwiz grand challenge: Answering visual questions from blind people}, | ||
author={Gurari, Danna and Li, Qing and Stangl, Abigale J and Guo, Anhong and Lin, Chi and Grauman, Kristen and Luo, Jiebo and Bigham, Jeffrey P}, | ||
booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition}, | ||
pages={3608--3617}, | ||
year={2018} | ||
} | ||
""", | ||
descriptive_stats={ | ||
"n_samples": {"test": 214354}, | ||
"avg_character_length": { | ||
"test": { | ||
"average_document_length": 0.0, | ||
"average_query_length": 0.0, | ||
"num_documents": 2143540, | ||
"num_queries": 214354, | ||
"average_relevant_docs_per_query": 10, | ||
} | ||
}, | ||
}, | ||
) |
186 changes: 186 additions & 0 deletions
186
...__clip-vit-base-patch32/3d74acf9a28c67741b2f4f2ea7635f0aaf6f0268/VizWizIT2TRetrieval.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 +1,186 @@ | ||
{ | ||
"dataset_revision": "723c6ffed034eb9dcbb44777e1a9ddf97c5ee28a", | ||
"evaluation_time": 32.021708726882935, | ||
"kg_co2_emissions": null, | ||
"mteb_version": "1.12.90", | ||
"scores": { | ||
"test": [ | ||
{ | ||
"cv_recall_at_1": 0.0294, | ||
"cv_recall_at_10": 0.07317, | ||
"cv_recall_at_100": 0.19819, | ||
"cv_recall_at_1000": 0.3737, | ||
"cv_recall_at_20": 0.09863, | ||
"cv_recall_at_3": 0.04608, | ||
"cv_recall_at_5": 0.05788, | ||
"hf_subset": "default", | ||
"languages": [ | ||
"eng-Latn" | ||
], | ||
"main_score": 0.01511, | ||
"map_at_1": 0.00301, | ||
"map_at_10": 0.00749, | ||
"map_at_100": 0.01081, | ||
"map_at_1000": 0.01176, | ||
"map_at_20": 0.0088, | ||
"map_at_3": 0.00507, | ||
"map_at_5": 0.00605, | ||
"mrr_at_1": 0.029404954850659876, | ||
"mrr_at_10": 0.041434029041114, | ||
"mrr_at_100": 0.045365167700329743, | ||
"mrr_at_1000": 0.04612327351299033, | ||
"mrr_at_20": 0.043187763329325876, | ||
"mrr_at_3": 0.036736898973527825, | ||
"mrr_at_5": 0.03942270587327312, | ||
"nauc_cv_recall_at_1000_diff1": 0.11813681061217512, | ||
"nauc_cv_recall_at_1000_max": -0.11894128215838488, | ||
"nauc_cv_recall_at_1000_std": 0.3782492530808885, | ||
"nauc_cv_recall_at_100_diff1": 0.11427439308010998, | ||
"nauc_cv_recall_at_100_max": -0.09531881188079029, | ||
"nauc_cv_recall_at_100_std": 0.2069207353440963, | ||
"nauc_cv_recall_at_10_diff1": 0.20419699826470403, | ||
"nauc_cv_recall_at_10_max": -0.11315602683152005, | ||
"nauc_cv_recall_at_10_std": 0.11079116202458081, | ||
"nauc_cv_recall_at_1_diff1": 0.3780410690211136, | ||
"nauc_cv_recall_at_1_max": -0.15996276050998642, | ||
"nauc_cv_recall_at_1_std": 0.12084151368405678, | ||
"nauc_cv_recall_at_20_diff1": 0.17482215533278464, | ||
"nauc_cv_recall_at_20_max": -0.09614325077967463, | ||
"nauc_cv_recall_at_20_std": 0.11622067910258244, | ||
"nauc_cv_recall_at_3_diff1": 0.3021499883385322, | ||
"nauc_cv_recall_at_3_max": -0.1409817803959759, | ||
"nauc_cv_recall_at_3_std": 0.12190544061690316, | ||
"nauc_cv_recall_at_5_diff1": 0.2422495719721318, | ||
"nauc_cv_recall_at_5_max": -0.11604909641982115, | ||
"nauc_cv_recall_at_5_std": 0.12525785371426437, | ||
"nauc_map_at_1000_diff1": 0.16429877210946825, | ||
"nauc_map_at_1000_max": -0.1295306253626044, | ||
"nauc_map_at_1000_std": 0.2062946969445036, | ||
"nauc_map_at_100_diff1": 0.1670600341719391, | ||
"nauc_map_at_100_max": -0.12982814569101003, | ||
"nauc_map_at_100_std": 0.18876332652966407, | ||
"nauc_map_at_10_diff1": 0.19138314983144086, | ||
"nauc_map_at_10_max": -0.12835016161145998, | ||
"nauc_map_at_10_std": 0.15183799317664892, | ||
"nauc_map_at_1_diff1": 0.3572077845896727, | ||
"nauc_map_at_1_max": -0.1591774719332963, | ||
"nauc_map_at_1_std": 0.1257863823021691, | ||
"nauc_map_at_20_diff1": 0.17047854383467648, | ||
"nauc_map_at_20_max": -0.12495405745802321, | ||
"nauc_map_at_20_std": 0.16271914594752626, | ||
"nauc_map_at_3_diff1": 0.25952802533398156, | ||
"nauc_map_at_3_max": -0.1473725924476619, | ||
"nauc_map_at_3_std": 0.13084611961877016, | ||
"nauc_map_at_5_diff1": 0.2102820356246173, | ||
"nauc_map_at_5_max": -0.13351667721655666, | ||
"nauc_map_at_5_std": 0.14460103788615644, | ||
"nauc_mrr_at_1000_diff1": 0.2882430552866343, | ||
"nauc_mrr_at_1000_max": -0.13678972890948193, | ||
"nauc_mrr_at_1000_std": 0.12898324685751192, | ||
"nauc_mrr_at_100_diff1": 0.2891333801979083, | ||
"nauc_mrr_at_100_max": -0.1371078360866261, | ||
"nauc_mrr_at_100_std": 0.12577922767556043, | ||
"nauc_mrr_at_10_diff1": 0.30678787633680227, | ||
"nauc_mrr_at_10_max": -0.1417415793576384, | ||
"nauc_mrr_at_10_std": 0.11840012286416857, | ||
"nauc_mrr_at_1_diff1": 0.3780410690211136, | ||
"nauc_mrr_at_1_max": -0.15996276050998642, | ||
"nauc_mrr_at_1_std": 0.12084151368405678, | ||
"nauc_mrr_at_20_diff1": 0.29802916652653827, | ||
"nauc_mrr_at_20_max": -0.1380992977602512, | ||
"nauc_mrr_at_20_std": 0.11913643771545089, | ||
"nauc_mrr_at_3_diff1": 0.34218208940138206, | ||
"nauc_mrr_at_3_max": -0.15259828173240922, | ||
"nauc_mrr_at_3_std": 0.11980199564671674, | ||
"nauc_mrr_at_5_diff1": 0.31906571381939547, | ||
"nauc_mrr_at_5_max": -0.14353131092973306, | ||
"nauc_mrr_at_5_std": 0.12118536859900152, | ||
"nauc_ndcg_at_1000_diff1": 0.1382546228263889, | ||
"nauc_ndcg_at_1000_max": -0.1186022001179963, | ||
"nauc_ndcg_at_1000_std": 0.29134546582965304, | ||
"nauc_ndcg_at_100_diff1": 0.1608084546142371, | ||
"nauc_ndcg_at_100_max": -0.1208877559776067, | ||
"nauc_ndcg_at_100_std": 0.20420457544308251, | ||
"nauc_ndcg_at_10_diff1": 0.21347296273381702, | ||
"nauc_ndcg_at_10_max": -0.12360631260617351, | ||
"nauc_ndcg_at_10_std": 0.13835357292918368, | ||
"nauc_ndcg_at_1_diff1": 0.35720778458967256, | ||
"nauc_ndcg_at_1_max": -0.15917747193329645, | ||
"nauc_ndcg_at_1_std": 0.12578638230216885, | ||
"nauc_ndcg_at_20_diff1": 0.189575288463788, | ||
"nauc_ndcg_at_20_max": -0.12047147115102609, | ||
"nauc_ndcg_at_20_std": 0.14823959023714925, | ||
"nauc_ndcg_at_3_diff1": 0.28072213819651337, | ||
"nauc_ndcg_at_3_max": -0.1493321151811889, | ||
"nauc_ndcg_at_3_std": 0.1313988281635346, | ||
"nauc_ndcg_at_5_diff1": 0.23487221921337986, | ||
"nauc_ndcg_at_5_max": -0.12854699731824779, | ||
"nauc_ndcg_at_5_std": 0.13838213366410212, | ||
"nauc_precision_at_1000_diff1": 0.09153318291899752, | ||
"nauc_precision_at_1000_max": -0.10712625620860772, | ||
"nauc_precision_at_1000_std": 0.3486808234128117, | ||
"nauc_precision_at_100_diff1": 0.11829783638630895, | ||
"nauc_precision_at_100_max": -0.11229273014278543, | ||
"nauc_precision_at_100_std": 0.2457874398625524, | ||
"nauc_precision_at_10_diff1": 0.1586640506586383, | ||
"nauc_precision_at_10_max": -0.10658265337446211, | ||
"nauc_precision_at_10_std": 0.1433984006213967, | ||
"nauc_precision_at_1_diff1": 0.35720778458967256, | ||
"nauc_precision_at_1_max": -0.15917747193329645, | ||
"nauc_precision_at_1_std": 0.12578638230216885, | ||
"nauc_precision_at_20_diff1": 0.13529505359357108, | ||
"nauc_precision_at_20_max": -0.10702832193841087, | ||
"nauc_precision_at_20_std": 0.16042770655667163, | ||
"nauc_precision_at_3_diff1": 0.2468790893829598, | ||
"nauc_precision_at_3_max": -0.1425487542554598, | ||
"nauc_precision_at_3_std": 0.1347458330420782, | ||
"nauc_precision_at_5_diff1": 0.17923534723711776, | ||
"nauc_precision_at_5_max": -0.10942897722644009, | ||
"nauc_precision_at_5_std": 0.1456878106837016, | ||
"nauc_recall_at_1000_diff1": 0.09153318291899752, | ||
"nauc_recall_at_1000_max": -0.1071262562086077, | ||
"nauc_recall_at_1000_std": 0.34868082341281187, | ||
"nauc_recall_at_100_diff1": 0.11829783638630909, | ||
"nauc_recall_at_100_max": -0.11229273014278539, | ||
"nauc_recall_at_100_std": 0.24578743986255244, | ||
"nauc_recall_at_10_diff1": 0.1586640506586383, | ||
"nauc_recall_at_10_max": -0.10658265337446211, | ||
"nauc_recall_at_10_std": 0.1433984006213967, | ||
"nauc_recall_at_1_diff1": 0.3572077845896727, | ||
"nauc_recall_at_1_max": -0.1591774719332963, | ||
"nauc_recall_at_1_std": 0.1257863823021691, | ||
"nauc_recall_at_20_diff1": 0.13529505359357108, | ||
"nauc_recall_at_20_max": -0.10702832193841087, | ||
"nauc_recall_at_20_std": 0.16042770655667163, | ||
"nauc_recall_at_3_diff1": 0.24687908938295963, | ||
"nauc_recall_at_3_max": -0.14254875425545974, | ||
"nauc_recall_at_3_std": 0.1347458330420781, | ||
"nauc_recall_at_5_diff1": 0.17923534723711776, | ||
"nauc_recall_at_5_max": -0.10942897722644009, | ||
"nauc_recall_at_5_std": 0.1456878106837016, | ||
"ndcg_at_1": 0.0301, | ||
"ndcg_at_10": 0.01511, | ||
"ndcg_at_100": 0.03019, | ||
"ndcg_at_1000": 0.04943, | ||
"ndcg_at_20": 0.01882, | ||
"ndcg_at_3": 0.02242, | ||
"ndcg_at_5": 0.01915, | ||
"precision_at_1": 0.0301, | ||
"precision_at_10": 0.01197, | ||
"precision_at_100": 0.00477, | ||
"precision_at_1000": 0.0012, | ||
"precision_at_20": 0.00935, | ||
"precision_at_3": 0.02022, | ||
"precision_at_5": 0.01635, | ||
"recall_at_1": 0.00301, | ||
"recall_at_10": 0.01197, | ||
"recall_at_100": 0.04767, | ||
"recall_at_1000": 0.11975, | ||
"recall_at_20": 0.01871, | ||
"recall_at_3": 0.00607, | ||
"recall_at_5": 0.00817 | ||
} | ||
] | ||
}, | ||
"task_name": "VizWizIT2TRetrieval" | ||
} |