VictorSanh HF staff commited on
Commit
5c14b2a
1 Parent(s): c954cad

include uda subset

Browse files
Files changed (1) hide show
  1. epic_kitchens_100.py +49 -24
epic_kitchens_100.py CHANGED
@@ -63,6 +63,7 @@ _URL_BASE = "https://raw.githubusercontent.com/epic-kitchens/epic-kitchens-100-a
63
  _VARIANTS = [
64
  "action_recognition", # This split is used by four challenges: Action Recognition, Weakly supervised action recognition, Action detection, Action anticipation
65
  "multi_instance_retrieval",
 
66
  ]
67
  class EpicKitchens100(datasets.GeneratorBasedBuilder):
68
  """Epic Kitchens"""
@@ -110,36 +111,59 @@ class EpicKitchens100(datasets.GeneratorBasedBuilder):
110
  "multi_instance_retrieval": {
111
  "train": os.path.join(_URL_BASE, "retrieval_annotations/EPIC_100_retrieval_train.csv"),
112
  "test": os.path.join(_URL_BASE, "retrieval_annotations/EPIC_100_retrieval_test.csv")
 
 
 
 
 
 
 
 
113
  }
114
  }
 
115
  files_path = dl_manager.download_and_extract(urls)
116
- splits = [
117
- datasets.SplitGenerator(
118
- name=datasets.Split.TRAIN,
119
- gen_kwargs={
120
- "annotations": files_path[self.config.name]["train"],
121
- "split": "train",
122
- },
123
- ),
124
- datasets.SplitGenerator(
125
- name=datasets.Split.TEST,
126
- gen_kwargs={
127
- "annotations": files_path[self.config.name]["test"],
128
- "split": "test",
129
- },
130
- ),
131
- ]
132
- if self.config.name == "action_recognition":
133
- splits.append(
 
 
 
 
134
  datasets.SplitGenerator(
135
- name=datasets.Split.VALIDATION,
136
  gen_kwargs={
137
- "annotations": files_path[self.config.name]["validation"],
138
- "split": "validation",
139
  },
140
  ),
141
- )
142
- return splits
 
 
 
 
 
 
 
 
 
 
143
 
144
  def _generate_examples(self, annotations, split):
145
  """This function returns the examples."""
@@ -148,7 +172,8 @@ class EpicKitchens100(datasets.GeneratorBasedBuilder):
148
  next(csv_reader) # Skip header
149
  for idx, row in enumerate(csv_reader):
150
  narration_id, participant_id, video_id, narration_timestamp, start_timestamp, stop_timestamp = row[:6]
151
- if split != "test":
 
152
  # The reason why it's jumping from 5 to 8 is that we are skipping `start_frame` and `stop_frame`
153
  # since we are not exposing the frames, but just the videos
154
  narration, verb, verb_class, noun, noun_class, all_nouns, all_noun_classes = row[8:15]
63
  _VARIANTS = [
64
  "action_recognition", # This split is used by four challenges: Action Recognition, Weakly supervised action recognition, Action detection, Action anticipation
65
  "multi_instance_retrieval",
66
+ "unsupervised_domain_adaptation",
67
  ]
68
  class EpicKitchens100(datasets.GeneratorBasedBuilder):
69
  """Epic Kitchens"""
111
  "multi_instance_retrieval": {
112
  "train": os.path.join(_URL_BASE, "retrieval_annotations/EPIC_100_retrieval_train.csv"),
113
  "test": os.path.join(_URL_BASE, "retrieval_annotations/EPIC_100_retrieval_test.csv")
114
+ },
115
+ "unsupervised_domain_adaptation": {
116
+ "source_train": os.path.join(_URL_BASE, "UDA_annotations/EPIC_100_uda_source_train.csv"),
117
+ "target_train": os.path.join(_URL_BASE, "UDA_annotations/EPIC_100_uda_target_train_timestamps.csv"),
118
+ "source_test": os.path.join(_URL_BASE, "UDA_annotations/EPIC_100_uda_source_test_timestamps.csv"),
119
+ "target_test": os.path.join(_URL_BASE, "UDA_annotations/EPIC_100_uda_target_test_timestamps.csv"),
120
+ "source_val": os.path.join(_URL_BASE, "UDA_annotations/EPIC_100_uda_source_val.csv"),
121
+ "target_val": os.path.join(_URL_BASE, "UDA_annotations/EPIC_100_uda_target_val.csv"),
122
  }
123
  }
124
+ # Download data for all splits once for all since they are tiny csv files
125
  files_path = dl_manager.download_and_extract(urls)
126
+
127
+ if self.config.name == "unsupervised_domain_adaptation":
128
+ splits = [
129
+ datasets.SplitGenerator(
130
+ name=datasets.Split(n_),
131
+ gen_kwargs={
132
+ "annotations": files_path[self.config.name][n_],
133
+ "split": n_,
134
+ },
135
+ )
136
+ for n_ in ["source_train", "target_train", "source_test", "target_test", "source_val", "target_val"]
137
+ ]
138
+ return splits
139
+ else:
140
+ splits = [
141
+ datasets.SplitGenerator(
142
+ name=datasets.Split.TRAIN,
143
+ gen_kwargs={
144
+ "annotations": files_path[self.config.name]["train"],
145
+ "split": "train",
146
+ },
147
+ ),
148
  datasets.SplitGenerator(
149
+ name=datasets.Split.TEST,
150
  gen_kwargs={
151
+ "annotations": files_path[self.config.name]["test"],
152
+ "split": "test",
153
  },
154
  ),
155
+ ]
156
+ if self.config.name == "action_recognition":
157
+ splits.append(
158
+ datasets.SplitGenerator(
159
+ name=datasets.Split.VALIDATION,
160
+ gen_kwargs={
161
+ "annotations": files_path[self.config.name]["validation"],
162
+ "split": "validation",
163
+ },
164
+ ),
165
+ )
166
+ return splits
167
 
168
  def _generate_examples(self, annotations, split):
169
  """This function returns the examples."""
172
  next(csv_reader) # Skip header
173
  for idx, row in enumerate(csv_reader):
174
  narration_id, participant_id, video_id, narration_timestamp, start_timestamp, stop_timestamp = row[:6]
175
+ if (self.config.name in ["action_recognition", "multi_instance_retrieval"] and split in ["train", "validation"]) or \
176
+ (self.config.name == "unsupervised_domain_adaptation" and split in ["source_train", "source_val", "target_val"]):
177
  # The reason why it's jumping from 5 to 8 is that we are skipping `start_frame` and `stop_frame`
178
  # since we are not exposing the frames, but just the videos
179
  narration, verb, verb_class, noun, noun_class, all_nouns, all_noun_classes = row[8:15]