Source code for nengo_dl.signals

"""
Represents and manages the internal simulation signals.
"""

from collections import defaultdict, OrderedDict, Mapping
import logging

from nengo.builder.signal import Signal
from nengo.exceptions import BuildError
import numpy as np
import tensorflow as tf


logger = logging.getLogger(__name__)


[docs]class TensorSignal: """ Represents a tensor as an indexed view into a base array. Parameters ---------- slices : tuple of tuple of int Start/stop indices of slices along the first axis of the base array, corresponding to the data for this signal. key : object Key mapping to the base array that contains the data for this signal. dtype : str dtype of the values represented by this signal. shape : tuple of int View shape of this signal (may differ from shape of base array). minibatch_size : int If not None then this signal contains a minibatch dimension with the given size. label : str Name for this signal, used to make debugging easier. """ def __init__(self, slices, key, dtype, shape, minibatch_size, label="TensorSignal"): # make sure slices are read-only slices = tuple(tuple(s) for s in slices) self._slices = slices self.key = key self.dtype = dtype self.shape = shape self.minibatch_size = minibatch_size self.label = label self.reset()
[docs] def reset(self): """ Reset cached Tensors. """ self._tf_shape = None self._tf_indices = None self._tf_indices_nd = None self._tf_slice = -1
@property def slices(self): """ The slices containing the data for this signal in the base array. """ return self._slices @slices.setter def slices(self, _): raise BuildError("Slices are read only") @property def ndim(self): """ The rank of this signal. """ return len(self.shape) def __repr__(self): return "TensorSignal(key=%s, shape=%s, label=%s)" % ( self.key, self.shape, self.label, )
[docs] def __getitem__(self, indices): """ Create a new TensorSignal representing a subset (slice or advanced indexing) of the indices of this TensorSignal. Parameters ---------- indices : slice or list of int The desired subset of the indices in this TensorSignal Returns ------- sig : `.signals.TensorSignal` A new TensorSignal representing the subset of this TensorSignal """ if indices is Ellipsis or indices is None: return self # figure out indices in new view source_indices = np.concatenate( [np.arange(start, stop, dtype=np.int32) for start, stop in self.slices] ) new_indices = source_indices[indices] # find all the entries that are not runs (not consecutive with the # previous entry) run_starts = np.empty(new_indices.shape[0], dtype=bool) run_starts[0] = True np.not_equal(new_indices[:-1] + 1, new_indices[1:], out=run_starts[1:]) # find run start/stop indices run_breaks = np.nonzero(run_starts)[0] starts = new_indices[run_breaks] stops = np.append(new_indices[run_breaks - 1][1:], new_indices[-1]) + 1 slices = tuple(zip(starts, stops)) return TensorSignal( slices, self.key, self.dtype, (len(new_indices),) + self.shape[1:], self.minibatch_size, label=self.label + ".slice", )
[docs] def reshape(self, shape): """ Create a new TensorSignal representing a reshaped view of the same data in this TensorSignal (size of data must remain unchanged). Parameters ---------- shape : tuple of int New shape for the signal (one dimension can be -1 to indicate an inferred dimension size, as in numpy) Returns ------- sig : `.signals.TensorSignal` New TensorSignal representing the same data as this signal but with the given shape """ # replace -1 with inferred dimension if shape.count(-1) > 1: raise BuildError("Only one inferred dimension allowed in reshape") elif shape.count(-1) == 1: n_elem = np.prod(self.shape) n_shape = int(np.prod([x for x in shape if x != -1])) if n_elem % n_shape != 0: raise BuildError("No valid length for inferred dimension") shape = tuple(x if x != -1 else n_elem // n_shape for x in shape) else: if np.prod(shape) != np.prod(self.shape): raise BuildError("Number of elements don't match in reshape") return TensorSignal( self.slices, self.key, self.dtype, shape, self.minibatch_size, label=self.label + ".reshape(%s)" % (shape,), )
@property def tf_shape(self): """ A ``tf.Tensor`` representing the shape of this signal. """ if self._tf_shape is None: self._tf_shape = tf.constant(self.full_shape, dtype=tf.int32) return self._tf_shape @property def tf_indices(self): """ A ``tf.Tensor`` representing the indices of this signal. """ if self._tf_indices is None: self._tf_indices = tf.concat( [tf.range(start, stop, dtype=tf.int32) for start, stop in self.slices], axis=0, ) return self._tf_indices @property def tf_indices_nd(self): """ A ``tf.Tensor`` representing the indices of this signal for use with e.g. ``scatter_nd``. """ if self._tf_indices_nd is None: if self.minibatched: self._tf_indices_nd = tf.stack( tf.meshgrid( tf.range(self.minibatch_size, dtype=tf.int32), self.tf_indices, indexing="ij", ), axis=-1, ) else: self._tf_indices_nd = tf.expand_dims(self.tf_indices, -1) return self._tf_indices_nd @property def tf_slice(self): """ A tuple of ``tf.Tensors`` representing the ``(start, stop, stride)`` slice within the base array containing the data for this signal. This can be used as a more efficient representation of `.TensorSignal.tf_indices`. """ if self._tf_slice == -1: if len(self.slices) == 1: start, stop = self.slices[0] if self.minibatched: # add full slice along first (batch) dimension start = [0, start] stop = [self.minibatch_size, stop] else: start = [start] stop = [stop] self._tf_slice = ( tf.constant(start), tf.constant(stop), ) else: self._tf_slice = None return self._tf_slice @property def full_shape(self): """Shape of the signal including the minibatch dimension.""" return ((self.minibatch_size,) + self.shape) if self.minibatched else self.shape @property def minibatched(self): """Whether or not this TensorSignal contains a minibatch dimension.""" return self.minibatch_size is not None
[docs]class SignalDict(Mapping): """ Handles the mapping from `~nengo.builder.Signal` to ``tf.Tensor``. Takes care of gather/scatter logic to read/write signals within the base arrays. Parameters ---------- dtype : str Floating point precision used in signals (e.g. "float32") minibatch_size : int Number of items in each minibatch """ def __init__(self, dtype, minibatch_size): self.dtype = tf.as_dtype(dtype) self.minibatch_size = minibatch_size self.sig_map = {} self.reset()
[docs] def reset(self): """ Reset build-specific data structures. These are data structures that are filled out during the TensorGraph build process (and therefore need to be re-initialized if we build the model again), as opposed to data that is constant for a given Nengo model. """ # these values will be re-generated whenever the model is rebuilt self.bases = OrderedDict() # reset TensorSignals for sig in self.sig_map.values(): sig.reset() # logging self.read_types = defaultdict(int) self.write_types = defaultdict(int)
[docs] def scatter(self, dst, val, mode="update"): """ Updates the base data corresponding to ``dst``. Parameters ---------- dst : `.TensorSignal` Signal indicating the data to be modified in base array val : ``tf.Tensor`` Update data (same shape as ``dst``, i.e. a dense array <= the size of the base array) mode : "update" or "inc" Overwrite/add the data at ``dst`` with ``val`` """ logger.debug("scatter") logger.debug("values %s", val) logger.debug("dst %s", dst) logger.debug("slices %s", dst.slices) logger.debug( "dst base %s", self.bases[dst.key] if dst.key in self.bases else None ) if val.dtype.is_floating and val.dtype.base_dtype != self.dtype: raise BuildError( "Tensor detected with wrong dtype (%s), should " "be %s." % (val.dtype.base_dtype, self.dtype) ) # should never be writing to a variable if isinstance(self.bases[dst.key], tf.Variable): raise BuildError("Scatter target should not be a Variable") if isinstance(self.bases[dst.key], tuple): # this is the first set operation for this signal assert mode == "update" base_shape = self.bases[dst.key] var = None else: self.bases[dst.key].shape.assert_is_fully_defined() base_shape = self.bases[dst.key].shape var = self.bases[dst.key] # align val shape with dst base shape val.shape.assert_is_fully_defined() dst_shape = list(base_shape) dst_shape[dst.minibatched] = dst.shape[0] if val.shape != dst_shape: val = tf.reshape(val, dst.tf_shape) if len(dst.slices) == 1 and val.shape == base_shape: if mode == "inc": result = var + val self.write_types["assign_add"] += 1 else: result = val self.write_types["assign"] += 1 elif mode == "inc": result = tf.tensor_scatter_nd_add(var, dst.tf_indices_nd, val) self.write_types["scatter_add"] += 1 else: if var is None: result = tf.scatter_nd(dst.tf_indices_nd, val, shape=base_shape) else: result = tf.tensor_scatter_nd_update(var, dst.tf_indices_nd, val) self.write_types["scatter_update"] += 1 self.bases[dst.key] = result logger.debug("new dst base %s", self.bases[dst.key])
[docs] def gather(self, src, force_copy=False): """ Fetches the data corresponding to ``src`` from the base array. Parameters ---------- src : `.TensorSignal` Signal indicating the data to be read from base array force_copy : bool If True, always perform a gather, not a slice (this forces a copy). Note that setting ``force_copy=False`` does not guarantee that a copy won't be performed. Returns ------- gathered : ``tf.Tensor`` Tensor object corresponding to a dense subset of data from the base array """ logger.debug("gather") logger.debug("src %s", src) logger.debug("slices %s", src.slices) logger.debug("src base %s", self.bases[src.key]) var = self.bases[src.key] assert isinstance(var, tf.Tensor) # we prefer to get the data via `strided_slice` or `identity` if # possible, as it is more efficient if force_copy or len(src.slices) > 1: result = tf.gather(var, src.tf_indices, axis=1 if src.minibatched else 0) self.read_types["gather"] += 1 elif src.slices[0][0] == 0 and src.slices[0][1] == var.shape[src.minibatched]: result = var self.read_types["identity"] += 1 else: result = tf.strided_slice(var, *src.tf_slice) self.read_types["strided_slice"] += 1 # reshape the data according to the shape set in `src`, if there is # one, otherwise keep the shape of the base array if result.shape != src.full_shape: result = tf.reshape(result, src.tf_shape) return result
[docs] def combine(self, sigs, label="Combine"): """ Combines several TensorSignals into one by concatenating along the first axis. Parameters ---------- sigs : list of `.TensorSignal` or `~nengo.builder.Signal` Signals to be combined label : str Name for combined signal (to help with debugging) Returns ------- sig : `.TensorSignal` New TensorSignal representing the concatenation of the data in ``sigs`` """ if len(sigs) == 0: return [] assert isinstance(sigs, (list, tuple)) assert isinstance(sigs[0], (Signal, TensorSignal)) sigs = [self[s] if isinstance(s, Signal) else s for s in sigs] # make sure all the signals have the same base # note: this also tells us that they have the same dtype and # minibatching key = sigs[0].key assert all(s.key == key for s in sigs) # make sure all signals have the same shape (except first axis, # which we're concatenating along); note, this can fail even if they # all have the same base, due to reshaping shape = (np.sum([s.shape[0] for s in sigs]),) + sigs[0].shape[1:] assert all(s.shape[1:] == shape[1:] for s in sigs) # combine slices from signals (possibly merging consecutive slices) combined_slices = [] for sig in sigs: if len(combined_slices) > 0 and combined_slices[-1][1] == sig.slices[0][0]: combined_slices = combined_slices[:-1] + [ (combined_slices[-1][0], sig.slices[0][1]) ] combined_slices.extend(sig.slices[1:]) else: combined_slices.extend(sig.slices) output = self.get_tensor_signal( combined_slices, key, sigs[0].dtype, shape, sigs[0].minibatched, label=label, ) return output
[docs] def get_tensor_signal( self, slices, key, dtype, shape, minibatched, signal=None, label="TensorSignal" ): """ Creates a new ``TensorSignal`` with the given properties. This should be used rather than instantiating a new TensorSignal directly, as it handles some extra book-keeping. Parameters ---------- slices : tuple of tuple of int Start/stop indices of slices along the first axis of the base array, corresponding to the data for this signal. key : object Key mapping to the base array that contains the data for this signal dtype : str dtype of the values represented by this signal shape : tuple of int View shape of this signal (may differ from shape of base array) minibatched : bool Whether or not this signal contains a minibatch dimension signal : `~nengo.builder.Signal` If not None, associate the new ``TensorSignal`` with the given ``Signal`` in the ``sig_map`` label : str Name for this signal, used to make debugging easier Returns ------- sig : `.TensorSignal` A new ``TensorSignal`` with the given properties """ tensor_sig = TensorSignal( slices, key, dtype, shape, self.minibatch_size if minibatched else None, label=label, ) if signal is not None: if signal.sparse: assert sum(stop - start for start, stop in slices) == signal.size assert shape == (signal.size,) else: assert sum(stop - start for start, stop in slices) == ( 1 if len(signal.shape) == 0 else signal.shape[0] ) assert signal.size == np.prod(shape) assert signal.minibatched == minibatched self[signal] = tensor_sig return tensor_sig
[docs] def op_constant(self, ops, op_sizes, attr, dtype, shape=(1, -1)): """ Creates a tensor representing the constant parameters of an op group. Parameters ---------- ops : list of object The operators for some merged group of ops op_sizes : list of int The number of constant elements in each op attr : str The attribute of the op that describes the constant parameter dtype : str Numeric type of the parameter shape : tuple of int Shape for returned constant (this will be ignored in the scalar case). The default adds an empty dimension for broadcasting along the batch axis. Returns ------- constant : ``tf.Tensor`` Tensor containing the values of ``attr`` for the given ops. This will be a scalar if all the ops have the same parameter value, or an array giving the parameter value for each element in each op. """ if not isinstance(dtype, tf.DType): dtype = tf.as_dtype(dtype) vals = [getattr(op, attr) for op in ops] if np.allclose(vals, vals[0]): return tf.constant(vals[0], dtype=dtype) assert len(op_sizes) == len(ops) v = np.zeros(sum(op_sizes), dtype=dtype.as_numpy_dtype) k = 0 for val, size in zip(vals, op_sizes): v[k : k + size] = val k += size if shape is not None: v = np.reshape(v, shape) return tf.constant(v, dtype=dtype)
def __getitem__(self, sig): return self.sig_map[sig] def __setitem__(self, sig, tensor_sig): self.sig_map[sig] = tensor_sig def __len__(self): return len(self.sig_map) def __iter__(self): return iter(self.sig_map) def __contains__(self, sig): return sig in self.sig_map