trainer: "sts"
model_name: "bertimbau-100m-europarl-eubookshop-ted2020-tatoeba-ct1-nli-gist10-sts-cosent20-v1"
base_model_name: "bertimbau-100m-europarl-eubookshop-ted2020-tatoeba-ct1-nli-gist10-v1"
loss_function: "cosent"
seed: 1  # best seed for this model from first 30 seeds
learning_rate: 1e-5
warmup_ratio: 0.1
weight_decay: 0.01
batch_size: 64
use_amp: True
epochs: 20
validations_per_epoch: 10

# HPs used by JRodrigues to train albertina-100m-portuguese-ptpt-encoder:
# learning_rate 1e-5
# lr_scheduler_type linear
# weight_decay 0.01
# per_device_train_batch_size 192
# gradient_accumulation_steps 1
# num_train_epochs 150
# num_warmup_steps 10000