Closes #425 | Add Dataloader MaXM

seacrowd/sea_datasets/bactrian_x/bactrian_x.py

@@ -0,0 +1,153 @@
+# coding=utf-8
+# Copyright 2022 The HuggingFace Datasets Authors and the current dataset script contributor.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import json
+from pathlib import Path
+from typing import Dict, List, Tuple
+
+import datasets
+
+from seacrowd.utils.configs import SEACrowdConfig
+from seacrowd.utils.constants import Tasks, Licenses, TASK_TO_SCHEMA, SCHEMA_TO_FEATURES
+
+_CITATION = """\
+@misc{li2023bactrianx,
+      title={Bactrian-X : A Multilingual Replicable Instruction-Following Model with Low-Rank Adaptation}, 
+      author={Haonan Li and Fajri Koto and Minghao Wu and Alham Fikri Aji and Timothy Baldwin},
+      year={2023},
+      eprint={2305.15011},
+      archivePrefix={arXiv},
+      primaryClass={cs.CL}
+}
+"""
+
+_DATASETNAME = "bactrian_x"
+
+_DESCRIPTION = """\
+The Bactrain-X dataset is a collection of 3.4M instruction-response pairs in 52
+languages, that are obtained by translating 67K English instructions (alpaca-52k
++ dolly-15k) into 51 languages using Google Translate API. The translated
+instructions are then fed to ChatGPT (gpt-3.5-turbo) to obtain its natural
+responses, resulting in 3.4M instruction-response pairs in 52 languages (52
+languages x 67k instances = 3.4M instances). Human evaluations were conducted to
+evaluate response quality for several languages, with those of interest to
+SEACrowd being Burmese and Tagalog.
+"""
+
+_HOMEPAGE = "https://github.com/mbzuai-nlp/Bactrian-X"
+
+_LANGUAGES = ["mya", "tgl", "ind", "khm", "tha", "vie"]
+
+_LICENSE = Licenses.CC_BY_NC_4_0.value
+
+_LOCAL = False
+
+_BASE_URL = "https://huggingface.co/datasets/MBZUAI/Bactrian-X/resolve/main/data/{subset}.json.gz?download=true"
+_SUBSETS = ["my", "tl", "id", "km", "th", "vi"]
+
+_SUPPORTED_TASKS = [Tasks.INSTRUCTION_TUNING]
+_SEACROWD_SCHEMA = f"seacrowd_{TASK_TO_SCHEMA[_SUPPORTED_TASKS[0]].lower()}"  # t2t
+
+_SOURCE_VERSION = "1.0.1"
+
+_SEACROWD_VERSION = "1.0.0"
+
+
+class BactrianXDataset(datasets.GeneratorBasedBuilder):
+    """A collection of translated instruction-response pairs, evaluated with ChatGPT and human."""
+
+    SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
+    SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
+
+    BUILDER_CONFIGS = []
+    for subset in _SUBSETS:
+        BUILDER_CONFIGS += [
+            SEACrowdConfig(
+                name=f"{_DATASETNAME}_{subset}_source",
+                version=SOURCE_VERSION,
+                description=f"{_DATASETNAME} {subset} source schema",
+                schema="source",
+                subset_id=subset,
+            ),
+            SEACrowdConfig(
+                name=f"{_DATASETNAME}_{subset}_{_SEACROWD_SCHEMA}",
+                version=SEACROWD_VERSION,
+                description=f"{_DATASETNAME} {subset} SEACrowd schema",
+                schema=_SEACROWD_SCHEMA,
+                subset_id=subset,
+            ),
+        ]
+
+    DEFAULT_CONFIG_NAME = f"{_DATASETNAME}_id_source"
+
+    def _info(self) -> datasets.DatasetInfo:
+        if self.config.schema == "source":
+            features = datasets.Features(
+                {
+                    "instruction": datasets.Value("string"),
+                    "input": datasets.Value("string"),
+                    "id": datasets.Value("string"),
+                    "output": datasets.Value("string"),
+                }
+            )
+        elif self.config.schema == _SEACROWD_SCHEMA:
+            features = SCHEMA_TO_FEATURES[
+                TASK_TO_SCHEMA[_SUPPORTED_TASKS[0]]
+            ]  # text2text_features
+
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=features,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
+        """Returns SplitGenerators."""
+        data_url = _BASE_URL.format(subset=self.config.name.split("_")[2])
+        data_path = Path(dl_manager.download_and_extract(data_url))
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={
+                    "data_path": data_path,
+                },
+            )
+        ]
+
+    def _generate_examples(self, data_path: Path) -> Tuple[int, Dict]:
+        """Yields examples as (key, example) tuples."""
+        with open(data_path, "r", encoding="utf-8") as file:
+            data = json.load(file)
+
+        if self.config.schema == "source":
+            for idx, example in enumerate(data):
+                yield idx, {
+                    "instruction": example["instruction"],
+                    "input": example["input"],
+                    "id": example["id"],
+                    "output": example["output"],
+                }
+        elif self.config.schema == _SEACROWD_SCHEMA:
+            for idx, example in enumerate(data):
+                yield idx, {
+                    "id": example["id"],
+                    "text_1": f"Instruction: {example['instruction']}\nInput: {example['input']}",
+                    "text_2": example["output"],
+                    "text_1_name": "instruction + input",
+                    "text_2_name": "output",
+                }

seacrowd/sea_datasets/maxm/maxm.py

@@ -0,0 +1,210 @@
+# coding=utf-8
+# Copyright 2022 The HuggingFace Datasets Authors and the current dataset script contributor.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import json
+from pathlib import Path
+from typing import Dict, List, Tuple
+
+import datasets
+
+from seacrowd.utils.configs import SEACrowdConfig
+from seacrowd.utils.constants import Tasks, Licenses, TASK_TO_SCHEMA, SCHEMA_TO_FEATURES
+
+_CITATION = """\
+@inproceedings{changpinyo-etal-2023-maxm,
+    title = "{M}a{XM}: Towards Multilingual Visual Question Answering",
+    author = "Changpinyo, Soravit  and
+      Xue, Linting  and
+      Yarom, Michal  and
+      Thapliyal, Ashish  and
+      Szpektor, Idan  and
+      Amelot, Julien  and
+      Chen, Xi  and
+      Soricut, Radu",
+    editor = "Bouamor, Houda  and
+      Pino, Juan  and
+      Bali, Kalika",
+    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2023",
+    month = dec,
+    year = "2023",
+    address = "Singapore",
+    publisher = "Association for Computational Linguistics",
+    url = "https://aclanthology.org/2023.findings-emnlp.176",
+    doi = "10.18653/v1/2023.findings-emnlp.176",
+    pages = "2667--2682",
+    abstract = "Visual Question Answering (VQA) has been primarily studied
+    through the lens of the English language. Yet, tackling VQA in other
+    languages in the same manner would require a considerable amount of
+    resources. In this paper, we propose scalable solutions to multilingual
+    visual question answering (mVQA), on both data and modeling fronts. We first
+    propose a translation-based framework to mVQA data generation that requires
+    much less human annotation efforts than the conventional approach of
+    directly collection questions and answers. Then, we apply our framework to
+    the multilingual captions in the Crossmodal-3600 dataset and develop an
+    efficient annotation protocol to create MaXM, a test-only VQA benchmark in 7
+    diverse languages. Finally, we develop a simple, lightweight, and effective
+    approach as well as benchmark state-of-the-art English and multilingual VQA
+    models. We hope that our benchmark encourages further research on mVQA.",
+}
+"""
+
+_DATASETNAME = "maxm"
+
+_DESCRIPTION = """\
+MaXM, a test-only VQA benchmark in 7 diverse languages, including Thai. The
+dataset is generated by first applying a translation-based framework to mVQA and
+then applying framework to the multilingual captions in the Crossmodal-3600
+dataset.
+"""
+
+_HOMEPAGE = "https://github.com/google-research-datasets/maxm"
+
+_LANGUAGES = ["tha"]
+
+_LICENSE = f"""{Licenses.OTHERS.value} | \
+The dataset may be freely used for any purpose, although acknowledgement of Google LLC ("Google") as the data source would be appreciated. 
+The dataset is provided "AS IS" without any warranty, express or implied.
+Google disclaims all liability for any damages, direct or indirect, resulting from the use of the dataset."""
+
+_LOCAL = False
+
+_URL = "https://storage.googleapis.com/maxm/maxm_v1_release.zip"
+_SUBSETS = ["regular", "yesno"]
+
+_SUPPORTED_TASKS = [Tasks.VISUAL_QUESTION_ANSWERING]
+_SEACROWD_SCHEMA = f"seacrowd_{TASK_TO_SCHEMA[_SUPPORTED_TASKS[0]].lower()}"  # imqa
+
+_SOURCE_VERSION = "1.0.0"
+
+_SEACROWD_VERSION = "1.0.0"
+
+
+class MaXMDataset(datasets.GeneratorBasedBuilder):
+    """A test-only VQA benchmark in 7 diverse languages, including Thai."""
+
+    SOURCE_VERSION = datasets.Version(_SOURCE_VERSION)
+    SEACROWD_VERSION = datasets.Version(_SEACROWD_VERSION)
+
+    BUILDER_CONFIGS = []
+    for subset in _SUBSETS:
+        BUILDER_CONFIGS += [
+            SEACrowdConfig(
+                name=f"{_DATASETNAME}_{subset}_source",
+                version=SOURCE_VERSION,
+                description=f"{_DATASETNAME} {subset} source schema",
+                schema="source",
+                subset_id=subset,
+            ),
+            SEACrowdConfig(
+                name=f"{_DATASETNAME}_{subset}_{_SEACROWD_SCHEMA}",
+                version=SEACROWD_VERSION,
+                description=f"{_DATASETNAME} {subset} SEACrowd schema",
+                schema=_SEACROWD_SCHEMA,
+                subset_id=subset,
+            ),
+        ]
+
+    DEFAULT_CONFIG_NAME = f"{_DATASETNAME}_regular_source"
+
+    def _info(self) -> datasets.DatasetInfo:
+        if self.config.schema == "source":
+            features = datasets.Features(
+                {
+                    "image_id": datasets.Value("string"),
+                    "image_url": datasets.Value("string"),
+                    "question_id": datasets.Value("string"),
+                    "question": datasets.Value("string"),
+                    "answers": datasets.Sequence(datasets.Value("string")),
+                    "processed_answers": datasets.Sequence(datasets.Value("string")),
+                    "is_collection": datasets.Value("bool"),
+                    "method": datasets.Value("string"),
+                }
+            )
+        elif self.config.schema == _SEACROWD_SCHEMA:
+            features = SCHEMA_TO_FEATURES[
+                TASK_TO_SCHEMA[_SUPPORTED_TASKS[0]]
+            ]  # imqa_features
+            features["meta"] = {
+                "processed_answers": datasets.Sequence(datasets.Value("string")),
+                "is_collection": datasets.Value("bool"),
+                "method": datasets.Value("string"),
+            }
+
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=features,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager: datasets.DownloadManager) -> List[datasets.SplitGenerator]:
+        """Returns SplitGenerators."""
+        data_path = Path(dl_manager.download_and_extract(_URL), "maxm_v1_release")
+        file_path = (
+            data_path
+            / f"maxm_v1_{'yesno_' if self.config.subset_id == 'yesno' else ''}th.json"
+        )
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={
+                    "filepath": file_path,
+                },
+            ),
+        ]
+
+    def _generate_examples(self, filepath: Path) -> Tuple[int, Dict]:
+        """Yields examples as (key, example) tuples."""
+        with open(filepath, "r", encoding="utf-8") as file:
+            data = json.load(file)
+
+        key = 0
+        data = data["annotations"]
+        if self.config.schema == "source":
+            for example in data:
+                for id, qa_pair in enumerate(example["qa_pairs"]):
+                    yield key, {
+                        "image_id": example["image_id"],
+                        "image_url": example["image_url"][id],
+                        "question_id": qa_pair["question_id"],
+                        "question": qa_pair["question"],
+                        "answers": qa_pair["answers"],
+                        "processed_answers": qa_pair["processed_answers"],
+                        "is_collection": qa_pair["is_collection"],
+                        "method": qa_pair["method"],
+                    }
+                    key += 1
+        elif self.config.schema == _SEACROWD_SCHEMA:
+            for example in data:
+                for id, qa_pair in enumerate(example["qa_pairs"]):
+                    yield key, {
+                        "id": str(key),
+                        "question_id": qa_pair["question_id"],
+                        "document_id": example["image_id"],
+                        "questions": [qa_pair["question"]],
+                        # "type": None,
+                        # "choices": None,
+                        # "context": None,
+                        "answer": qa_pair["answers"],
+                        "image_paths": [example["image_url"][id]],
+                        "meta": {
+                            "processed_answers": qa_pair["processed_answers"],
+                            "is_collection": qa_pair["is_collection"],
+                            "method": qa_pair["method"],
+                        },
+                    }
+                    key += 1