MCPcopy Create free account
hub / github.com/davisking/dlib / finalize_act_output

Function finalize_act_output

dlib/cuda/tensor_tools.cpp:1646–1666  ·  view source on GitHub ↗

Source from the content-addressed store, hash-verified

1644 }
1645
1646 void finalize_act_output(
1647 resizable_tensor& output,
1648 const tensor& input_data,
1649 const tensor& remainders,
1650 resizable_tensor& effective_weights,
1651 long batch_size,
1652 long seq_len,
1653 long d_model,
1654 long num_channels
1655 )
1656 {
1657 IF_DLIB_USE_CUDA(
1658 cuda::finalize_act_output(output, input_data, remainders, effective_weights,
1659 batch_size, seq_len, d_model, num_channels);
1660 )
1661
1662 IF_DLIB_NOT_USE_CUDA(
1663 cpu::finalize_act_output(output, input_data, remainders, effective_weights,
1664 batch_size, seq_len, d_model, num_channels);
1665 )
1666 }
1667
1668 void apply_act_depth_scaling(
1669 tensor& gradients,

Callers

nothing calls this directly

Calls

no outgoing calls

Tested by

no test coverage detected