Add support for multiprocessing for visual genome

pmd.py

@@ -17,7 +17,9 @@ import json
 import re
 import sqlite3
 from abc import abstractmethod
+from functools import partial
 from hashlib import md5
+from multiprocessing import Pool
 from pathlib import Path
 from typing import Any, Dict, List, Optional
 
@@ -26,7 +28,7 @@ from urllib.parse import unquote_plus
 
 # TODO: @thomasw21
 import datasets
-from datasets import load_dataset
+from datasets import load_dataset, Dataset
 from langdetect import detect
 
 _CITATION = """"""
@@ -84,11 +86,13 @@ class DatasetsLoader(BaseLoader):
         dataset_name: str,
         config_name: Optional[str],
         split: str,
+        num_proc: int,
         batch_size: int = 1000,
     ):
         super(DatasetsLoader, self).__init__(source=dataset_name, split=split)
         self.dataset_name = dataset_name
         self.config_name = config_name
+        self.num_proc = num_proc
         self.batch_size = batch_size
 
     @abstractmethod
@@ -105,16 +109,38 @@ class DatasetsLoader(BaseLoader):
             for i in range(batch_size)
         ]
 
+    def _generate_rows_casted_pmd_features(self, dset: Dataset, batch_start: int):
+        dataset_size = len(dset)
+        batch_end = min(batch_start + self.batch_size, dataset_size)
+        batch = dset[batch_start:batch_end]
+        rows = self.convert_batch_to_list_of_rows(batch)
+
+        rows_casted_pmd_features = []
+        for row in rows:
+            rows_casted_pmd_features += self.cast_to_pmd_features(row)
+        return rows_casted_pmd_features
+
+
     def _generate_examples(self):
         dataset = load_dataset(self.dataset_name, self.config_name, split=self.split)
         dataset_size = len(dataset)
+
         # load batches and yield individual rows
-        for batch_start in range(0, dataset_size, self.batch_size):
-            batch_end = min(batch_start + self.batch_size, dataset_size)
-            batch = dataset[batch_start:batch_end]
-            rows = self.convert_batch_to_list_of_rows(batch)
-            for row in rows:
-                rows_casted_pmd_features = self.cast_to_pmd_features(row)
+        if self.num_proc == 1:
+            for batch_start in range(0, dataset_size, self.batch_size):
+                rows_casted_pmd_features = self._generate_rows_casted_pmd_features(dset=dataset, batch_start=batch_start)
+                for row_casted_pmd_features in rows_casted_pmd_features:
+                    yield row_casted_pmd_features
+
+        # Parallel version
+        else:
+            assert self.num_proc > 1
+            with Pool(self.num_proc) as pool:
+                rows_casted_pmd_features = pool.imap(
+                    partial(self._generate_rows_casted_pmd_features, dset=dataset),
+                    range(0, dataset_size, self.batch_size),
+                    chunksize=10
+                )
                 for row_casted_pmd_features in rows_casted_pmd_features:
                     yield row_casted_pmd_features
 
@@ -211,6 +237,7 @@ class SBULoader(DatasetsLoader):
             config_name=None,
             split=split,
             batch_size=batch_size,
+            num_proc=1
         )
 
     def cast_to_pmd_features(self, row: Dict) -> List[Dict[str, Any]]:
@@ -426,12 +453,13 @@ class LocalizedNarrativesADE20kLoader(BaseLoaderWithDLManager):
 
 
 class VisualGenomeLoader(DatasetsLoader):
-    def __init__(self, split: str, batch_size: int = 1000):
+    def __init__(self, split: str, num_proc: int, batch_size: int = 1000):
         super(VisualGenomeLoader, self).__init__(
             dataset_name="visual_genome",
             config_name="region_descriptions_v1.2.0",
             split=split,
             batch_size=batch_size,
+            num_proc=num_proc
         )
 
     def cast_to_pmd_features(self, row: Dict) -> List[Dict[str, Any]]:
@@ -468,6 +496,7 @@ class WITLoader(DatasetsLoader):
             config_name=None,
             split=split,
             batch_size=batch_size,
+            num_proc=1
         )
 
     def cast_to_pmd_features(self, row: Dict) -> List[Dict[str, Any]]:
@@ -502,6 +531,7 @@ class ConceptualCaptions(DatasetsLoader):
             config_name="unlabeled",
             split=split,
             batch_size=batch_size,
+            num_proc=1
         )
 
     def cast_to_pmd_features(self, row: Dict) -> List[Dict[str, Any]]:
@@ -528,6 +558,7 @@ class Conceptual12MLoader(DatasetsLoader):
             config_name=None,
             split=split,
             batch_size=batch_size,
+            num_proc=1
         )
 
     def cast_to_pmd_features(self, row):
@@ -554,6 +585,7 @@ class RedCapsLoader(DatasetsLoader):
             config_name="all",
             split=split,
             batch_size=batch_size,
+            num_proc=1
         )
 
     def cast_to_pmd_features(self, row: Dict) -> List[Dict[str, Any]]:
@@ -732,15 +764,22 @@ class PMDConfig(datasets.BuilderConfig):
 
     def __init__(
         self,
+        num_proc: Optional[int] = None,
         datasets_batch_size: int = 1000,
         sqlite3_batch_size: int = 10_000,
         **kwargs,
     ):
+        if num_proc is None:
+            # We disable multiprocessing.
+            num_proc = 1
         super(PMDConfig, self).__init__(**kwargs)
         # determines how much we can load
         self.datasets_batch_size = datasets_batch_size
         self.sqlite3_batch_size = sqlite3_batch_size
 
+        # Some datasets should be loaded via multiprocessing.
+        self.num_proc = num_proc
+
 
 class PMD(datasets.GeneratorBasedBuilder):
     """Builder for Open Images subset of PMD."""
@@ -762,48 +801,49 @@ class PMD(datasets.GeneratorBasedBuilder):
                 name=split_name,
                 gen_kwargs={
                     "loaders": [
-                        COCOloader(dl_manager=dl_manager, split=split_name),
-                        SBULoader(
-                            split=split_name,
-                            batch_size=self.config.datasets_batch_size,
-                        ),
-                        LocalizedNarrativesOpenImagesLoader(
-                            dl_manager=dl_manager, split=split_name
-                        ),
-                        LocalizedNarrativesCOCOLoader(
-                            dl_manager=dl_manager, split=split_name
-                        ),
-                        LocalizedNarrativesFlickr30kLoader(
-                            dl_manager=dl_manager, split=split_name
-                        ),
-                        LocalizedNarrativesADE20kLoader(
-                            dl_manager=dl_manager, split=split_name
-                        ),
-                        ConceptualCaptions(
-                            split=split_name,
-                            batch_size=self.config.datasets_batch_size,
-                        ),
+                        # COCOloader(dl_manager=dl_manager, split=split_name),
+                        # SBULoader(
+                        #     split=split_name,
+                        #     batch_size=self.config.datasets_batch_size,
+                        # ),
+                        # LocalizedNarrativesOpenImagesLoader(
+                        #     dl_manager=dl_manager, split=split_name
+                        # ),
+                        # LocalizedNarrativesCOCOLoader(
+                        #     dl_manager=dl_manager, split=split_name
+                        # ),
+                        # LocalizedNarrativesFlickr30kLoader(
+                        #     dl_manager=dl_manager, split=split_name
+                        # ),
+                        # LocalizedNarrativesADE20kLoader(
+                        #     dl_manager=dl_manager, split=split_name
+                        # ),
+                        # ConceptualCaptions(
+                        #     split=split_name,
+                        #     batch_size=self.config.datasets_batch_size,
+                        # ),
                         VisualGenomeLoader(
                             split=split_name,
                             batch_size=self.config.datasets_batch_size,
+                            num_proc=self.config.num_proc
                         ),
-                        WITLoader(
-                            split=split_name,
-                            batch_size=self.config.datasets_batch_size,
-                        ),
-                        Conceptual12MLoader(
-                            split=split_name,
-                            batch_size=self.config.datasets_batch_size,
-                        ),
-                        RedCapsLoader(
-                            split=split_name,
-                            batch_size=self.config.datasets_batch_size,
-                        ),
-                        YFCC100MLoader(
-                            dl_manager=dl_manager,
-                            split=split_name,
-                            batch_size=self.config.sqlite3_batch_size,
-                        ),
+                        # WITLoader(
+                        #     split=split_name,
+                        #     batch_size=self.config.datasets_batch_size,
+                        # ),
+                        # Conceptual12MLoader(
+                        #     split=split_name,
+                        #     batch_size=self.config.datasets_batch_size,
+                        # ),
+                        # RedCapsLoader(
+                        #     split=split_name,
+                        #     batch_size=self.config.datasets_batch_size,
+                        # ),
+                        # YFCC100MLoader(
+                        #     dl_manager=dl_manager,
+                        #     split=split_name,
+                        #     batch_size=self.config.sqlite3_batch_size,
+                        # ),
                     ]
                 },
             )