def build_processors(self):
print('Tiki : Initializing : Building - Text Processors')
with open('/final/data/pythia.yaml') as f:
config = yaml.load(f, Loader=yaml.FullLoader)
config = ConfigNode(config)
config.training_parameters.evalai_inference = True # Remove warning
registry.register('config', config)
self.config = config
vqa_config = config.task_attributes.vqa.dataset_attributes.vqa2
text_processor_config = vqa_config.processors.text_processor
answer_processor_config = vqa_config.processors.answer_processor
text_processor_config.params.vocab.vocab_file = '/final/data/vocabulary_100k.txt'
answer_processor_config.params.vocab_file = '/final/data/answers_vqa.txt'
self.text_processor = VocabProcessor(text_processor_config.params)
self.answer_processor = VQAAnswerProcessor(
answer_processor_config.params)
registry.register('vqa2_text_processor', self.text_processor)
registry.register('vqa2_answer_processor', self.answer_processor)
registry.register('vqa2_num_final_outputs',
self.answer_processor.get_vocab_size())
def _init_processors(self):
with open(os.path.join(BASE_VQA_DIR_PATH, "model_data/pythia.yaml")) as f:
config = yaml.load(f)
config = ConfigNode(config)
# Remove warning
config.training_parameters.evalai_inference = True
registry.register("config", config)
self.config = config
vqa_config = config.task_attributes.vqa.dataset_attributes.vqa2
text_processor_config = vqa_config.processors.text_processor
answer_processor_config = vqa_config.processors.answer_processor
text_processor_config.params.vocab.vocab_file = os.path.join(
BASE_VQA_DIR_PATH, "model_data/vocabulary_100k.txt"
)
answer_processor_config.params.vocab_file = os.path.join(
BASE_VQA_DIR_PATH, "model_data/answers_vqa.txt"
)
# Add preprocessor as that will needed when we are getting questions from user
self.text_processor = VocabProcessor(text_processor_config.params)
self.answer_processor = VQAAnswerProcessor(answer_processor_config.params)
registry.register("vqa2_text_processor", self.text_processor)
registry.register("vqa2_answer_processor", self.answer_processor)
registry.register(
"vqa2_num_final_outputs", self.answer_processor.get_vocab_size()
)
def __init__(self, use_constrained=False):
super(PythiaCaptioner, self).__init__()
# load configuration file
with open(config_file) as f:
config = yaml.load(f)
config = ConfigNode(config)
self.use_constrained = use_constrained
# TODO: not sure what these two lines really means
config.training_parameters.evalai_inference = True
registry.register("config", config)
self.config = config
captioning_config = config.task_attributes.captioning.dataset_attributes.coco
text_processor_config = captioning_config.processors.text_processor
caption_processor_config = captioning_config.processors.caption_processor
text_processor_config.params.vocab.vocab_file = vocab_file
caption_processor_config.params.vocab.vocab_file = vocab_file
self.text_processor = VocabProcessor(text_processor_config.params)
self.caption_processor = CaptionProcessor(
caption_processor_config.params)
registry.register("coco_text_processor", self.text_processor)
registry.register("coco_caption_processor", self.caption_processor)
self.model = self._build_model()
def __init__(self, use_constrained=False):
super(PythiaCaptioner, self).__init__()
# load configuration file
with open(config_file) as f:
config = yaml.load(f)
config = ConfigNode(config)
self.use_constrained = use_constrained
# the following blocks of code read some configuration
# parameter in Pythia
config.training_parameters.evalai_inference = True
registry.register("config", config)
self.config = config
captioning_config = config.task_attributes.captioning.dataset_attributes.coco
text_processor_config = captioning_config.processors.text_processor
caption_processor_config = captioning_config.processors.caption_processor
# text_processor and caption_processor are used to pre-process the text
text_processor_config.params.vocab.vocab_file = vocab_file
caption_processor_config.params.vocab.vocab_file = vocab_file
self.text_processor = VocabProcessor(text_processor_config.params)
self.caption_processor = CaptionProcessor(
caption_processor_config.params)
registry.register("coco_text_processor", self.text_processor)
registry.register("coco_caption_processor", self.caption_processor)
self.model = self._build_model()
def _init_processors(self):
with open(model_yaml) as f:
config = yaml.load(f)
config = ConfigNode(config)
# Remove warning
config.training_parameters.evalai_inference = True
registry.register("config", config)
self.config = config
captioning_config = config.task_attributes.captioning.dataset_attributes.coco
# captioning_config = config.task_attributes.captioning.dataset_attributes.youcookII
text_processor_config = captioning_config.processors.text_processor
caption_processor_config = captioning_config.processors.caption_processor
# print("DEBUG captioning_config:", captioning_config)
# print("DEBUG text_processor_config:", text_processor_config)
# print("DEBUG caption_processor_config:", caption_processor_config)
text_processor_config.params.vocab.vocab_file = "content/model_data/vocabulary_captioning_thresh5.txt"
caption_processor_config.params.vocab.vocab_file = "content/model_data/vocabulary_captioning_thresh5.txt"
self.text_processor = VocabProcessor(text_processor_config.params)
self.caption_processor = CaptionProcessor(caption_processor_config.params)
# print("DEBUG text_processor:", self.text_processor)
# print("DEBUG caption_processor:", self.caption_processor)
registry.register("coco_text_processor", self.text_processor)
registry.register("coco_caption_processor", self.caption_processor)
def __init__(self):
config_file = 'model_data/butd.yaml'
vocab_file = 'model_data/vocabulary_captioning_thresh5.txt'
with open(config_file) as f:
config = yaml.load(f)
config = ConfigNode(config)
captioning_config = config.task_attributes.captioning.dataset_attributes.coco
text_processor_config = captioning_config.processors.text_processor
text_processor_config.params.vocab.vocab_file = vocab_file
text_processor = VocabProcessor(text_processor_config.params)
self.vocab = text_processor.vocab
def init_processors(caption_config: Dict, butd_config: Dict):
"""Build the caption and text processors.
"""
captioning_config = butd_config.task_attributes.captioning \
.dataset_attributes.coco
text_processor_config = captioning_config.processors.text_processor
caption_processor_config = captioning_config.processors \
.caption_processor
vocab_file_path = caption_config["text_caption_processor_vocab_txt"]
text_processor_config.params.vocab.vocab_file = vocab_file_path
caption_processor_config.params.vocab.vocab_file = vocab_file_path
text_processor = VocabProcessor(text_processor_config.params)
caption_processor = CaptionProcessor(caption_processor_config.params)
registry.register("coco_text_processor", text_processor)
registry.register("coco_caption_processor", caption_processor)
return caption_processor, text_processor
def _init_processors(self):
with open("model_data/butd.yaml") as f:
config = yaml.load(f)
config = ConfigNode(config)
config.training_parameters.evalai_inference = True
registry.register("config", config)
self.config = config
captioning_config = config.task_attributes.captioning.dataset_attributes.coco
text_processor_config = captioning_config.processors.text_processor
caption_processor_config = captioning_config.processors.caption_processor
text_processor_config.params.vocab.vocab_file = "model_data/vocabulary_captioning_thresh5.txt"
caption_processor_config.params.vocab.vocab_file = "model_data/vocabulary_captioning_thresh5.txt"
self.text_processor = VocabProcessor(text_processor_config.params)
self.caption_processor = CaptionProcessor(caption_processor_config.params)
registry.register("coco_text_processor", self.text_processor)
registry.register("coco_caption_processor", self.caption_processor)
#!/usr/bin/env python3
import yaml
from pythia.utils.configuration import ConfigNode
from pythia.tasks.processors import VocabProcessor, VQAAnswerProcessor
from torchvision import models
if __name__ == '__main__':
resnet152 = models.resnet152(pretrained=True)
with open('/final/data/pythia.yaml') as f:
config = yaml.load(f, Loader=yaml.FullLoader)
config = ConfigNode(config)
vqa_config = config.task_attributes.vqa.dataset_attributes.vqa2
text_processor_config = vqa_config.processors.text_processor
answer_processor_config = vqa_config.processors.answer_processor
text_processor_config.params.vocab.vocab_file = '/final/data/vocabulary_100k.txt'
answer_processor_config.params.vocab_file = '/final/data/answers_vqa.txt'
text_processor = VocabProcessor(text_processor_config.params)
answer_processor = VQAAnswerProcessor(answer_processor_config.params)