Torch Utils

• checkpoint_helper
  • build_checkpoint_helper
  • CheckpointHelper
  • CountVar
  • auto_checkpoint
• data_helper
  • to_device
  • to_dtype
  • to_tensor
  • to_ndarray
  • to_list
  • tensor_to_list
  • same_shape
  • build_log_buffer
  • CudaFetcher
  • get_tensor_data
• distribution
  • Pd
  • CategoricalPd
  • CategoricalPdPytorch
• metric
  • levenshtein_distance
  • hamming_distance
• nn_test_helper
  • is_differentiable
• optimizer_helper
  • Adam
  • RMSprop
• loss.cross_entropy_loss
  • LabelSmoothCELoss
  • SoftFocalLoss
  • build_ce_criterion
• loss.multi_logits_loss
  • MultiLogitsLoss
• network.activation
  • GLU
  • build_activation
• network.nn_module
  • weight_init
  • sequential_pack
  • conv1d_block
  • conv2d_block
  • deconv2d_block
  • fc_block
  • MLP
  • one_hot
  • binary_encode
  • noise_block
  • ChannelShuffle
  • NearestUpsample
  • BilinearUpsample
  • NoiseLinearLayer
• network.normalization
  • build_normalization
• network.res_block
  • ResBlock
  • ResFCBlock
• network.rnn
  • LSTMForwardWrapper
  • LSTM
  • PytorchLSTM
  • get_lstm
• network.scatter_connection
  • ScatterConnection
• network.soft_argmax
  • SoftArgmax
• network.transformer
  • Attention
  • TransformerLayer
  • Transformer