n_pair_mc_cars196.yaml

num_epochs: 20
num_batches_per_epoch: 500
batch_size: 120
out_dim: 128
learning_rate: !!float 7e-5
loss_l2_reg: !!float 3e-3
l2_weight_decay: !!float 5e-3
normalize_output: False
optimizer: 'Adam'
crop_size: 224
distance_type: 'euclidean'
dataset: 'cars196'
method: 'n_pairs_mc'  # sampling method for batch construction