configs/cifar100/default_cifar100.yml
# Training parameters:
training:
task:
type: CIFAR100
batch_size: &b 1024
use_train_data: True
# Use sampler that operates on a subset.
dataloader:
num_workers: 10
# shuffle: False
sampler:
type: SubsetRandomSampler
indices: [0, 45000]
# optimizer parameters:
optimizer:
type: Adam
lr: 0.001
# settings parameters
terminal_conditions:
loss_stop_threshold: 0.05
early_stop_validations: -1
episode_limit: 10000
epoch_limit: 10
# Validation parameters:
validation:
#partial_validation_interval: 100
task:
type: CIFAR100
batch_size: *b
use_train_data: True # True because we are splitting the training set to: validation and training
#resize: [32, 32]
# Use sampler that operates on a subset.
sampler:
type: SubsetRandomSampler
indices: [45000, 50000]
# Testing parameters:
test:
task:
type: MNIST
batch_size: *b
use_train_data: False
#resize: [32, 32]
pipeline:
disable: image_viewer
# Loss
nllloss:
type: NLLLoss
priority: 10.0
streams:
targets: fine_targets
# Statistics.
batch_size:
priority: 100.0
type: BatchSizeStatistics
streams:
targets: fine_targets
accuracy:
priority: 100.1
type: AccuracyStatistics
streams:
targets: fine_targets
precision_recall:
priority: 100.2
type: PrecisionRecallStatistics
use_word_mappings: True
#show_class_scores: True
globals:
word_mappings: fine_label_word_mappings
streams:
targets: fine_targets
answer_decoder:
priority: 100.3
type: WordDecoder
import_word_mappings_from_globals: True
globals:
word_mappings: fine_label_word_mappings
streams:
inputs: predictions
outputs: answers
stream_viewer:
priority: 100.4
type: StreamViewer
input_streams: coarse_targets, coarse_labels, fine_targets, fine_labels, answers
image_viewer:
priority: 100.5
type: ImageViewer
streams:
images: inputs
labels: fine_labels
answers: coarse_labels