DebugHelpers

This file is going to be imported by Debug.debug_shell() and available as interactive commands.

DebugHelpers.find_obj_in_stack(cls, stack=None, all_threads=True)[source]
DebugHelpers.get_device()[source]
Return type:Device.Device
DebugHelpers.compute(var, trainnet=True)[source]
Parameters:
  • var (theano.Variable) – variable which we should compute the value of
  • trainnet (bool) – whether to make givens based on dev.trainnet or dev.testnet
Returns:

the computed value

Return type:

numpy.ndarray

This expects to calculate some value of the trainnet or testnet of the current Device.

class DebugHelpers.DebugNn(filename)[source]
compile_forwarder()[source]
forward(data, output_index=0)[source]
class DebugHelpers.SimpleHdf(filename)[source]
get_seq_tags()[source]
get_data(seq_idx)[source]
get_targets(seq_idx)[source]
get_data_dict(seq_idx)[source]