Mercurial > pylearn
view nnet_ops.py @ 39:c682c6e9bf93
Minor edits
author | bengioy@esprit.iro.umontreal.ca |
---|---|
date | Thu, 24 Apr 2008 14:46:10 -0400 |
parents | 1b152f46ad0c |
children | 810a8e3c85e1 |
line wrap: on
line source
import theano from theano import tensor, gof, scalar import numpy class ScalarSigmoid(scalar.UnaryScalarOp): def impl(self, x): return 1.0 / (1 + numpy.exp(-x)) def grad(self, (x,), (gz,)): return gz * scalar_sigmoid(x) * (1.0 - scalar_sigmoid(x)), def c_foreach(self, (x,), (z,)): return "%(z)s = 1.0 / (1 + exp(-%(x)s));" % locals() scalar_sigmoid = gof.op.constructor(ScalarSigmoid) Sigmoid, sigmoid, SigmoidInplace, sigmoid_inplace \ = theano.tensor.broadcast(ScalarSigmoid, 'Sigmoid') class CrossentropySoftmax1Hot(gof.op.Op): """A special compound Op for the output of neural-net classifiers. This Op has two outputs: - KL(softmax(x), y) - softmax(x) x[i] is assumed to be a dense vector softmax(x[i]) is the i'th distribution over len(x[i]) options y[i] is an integer index, encoding a 1-hot distribution """ nin=2 nout=2 def __init__(self, x, b, y_idx, **kwargs): x = tensor._as_tensor(x) b = tensor._as_tensor(b) y_idx = tensor._as_tensor(y_idx) if len(x.broadcastable) != 2 \ or x.dtype not in ['float32', 'float64']: raise ValueError('x must be 2-d tensor of floats') if len(b.broadcastable) != 1 \ or x.dtype not in ['float32', 'float64']: raise ValueError('x must be 1-d tensor of floats') if len(y_idx.broadcastable) != 1 \ or y_idx.dtype not in ['int32', 'int64']: raise ValueError('x must be 1-d tensor of ints') # TODO: Is this correct? It used to be y, not y_idx nll = tensor.Tensor(x.dtype, y_idx.broadcastable) # nll = Tensor(x.dtype, y.broadcastable) sm = tensor.Tensor(x.dtype, x.broadcastable) self.inputs = [x, b, y_idx] self.outputs = [nll, sm] def perform(self): x, b, y_idx = [i.data for i in self.inputs] if b.shape[0] != x.shape[1]: raise ValueError('b must have same shape as x[0]') sm = numpy.zeros_like(x) # softmax nll = numpy.zeros(x.shape[0]) #nll(y | softmax(x)) for i in xrange(sm.shape[0]): row = x[i] + b sm[i] = numpy.exp(row - numpy.max(row)) #softmax sm[i] *= 1.0 / numpy.sum(sm[i]) #vector scale nll[i] = -numpy.log( sm[i, y_idx[i]]) #cross-entropy self.outputs[0].data = nll self.outputs[1].data = sm def grad(self, (x, b, y_idx), (g_nll, g_sm)): if g_sm is not None: raise NotImplementedError() nll, sm = crossentropy_softmax_1hot(x, b, y_idx) dx = CrossentropySoftmax1HotDx(g_nll, sm, y_idx).outputs[0] db = tensor.Sum(dx, axis = [0]).outputs[0] return dx, db, None def c_validate_cleanup(self, (x, b, y_idx), (nll, sm), sub): """Not sure...""" return "" def c_support_code(self): return """ """ def c_code(self, (x, b, y_idx), (nll, sm), sub): # this implementation was lifted from # /u/bergstrj/cvs/bergstrj/src/feb07/nn.cxx #TODO: put this into a templated function, in the support code #TODO: declare the max of each row as an Op output #TODO: set error messages for failures in this code return """ npy_intp* Nx = %(x)s->dimensions; if (%(x)s->nd != 2) { %(fail)s } if (%(b)s->nd != 1) { %(fail)s } if (%(y_idx)s->nd != 1) { %(fail)s } if (%(x)s->descr->type_num != PyArray_DOUBLE) { %(fail)s} if (%(b)s->descr->type_num != PyArray_DOUBLE) { %(fail)s} if (%(y_idx)s->descr->type_num != PyArray_INT64) { %(fail)s} %(nll)s = (PyArrayObject*)PyArray_SimpleNew(1, PyArray_DIMS(%(y_idx)s), type_num_%(x)s); if(!%(nll)s){%(fail)s} %(sm)s = (PyArrayObject*)PyArray_SimpleNew(2, PyArray_DIMS(%(x)s), type_num_%(x)s); if(!%(sm)s) { // The normal cleanup code will take care of %(nll)s // Py_XDECREF(%(nll)s); %(nll)s=NULL; %(fail)s } if (%(x)s->dimensions[1] != %(b)s->dimensions[0]) {%(fail)s} if (%(sm)s->dimensions[0] != %(x)s->dimensions[0]) {%(fail)s} if (%(sm)s->dimensions[1] != %(x)s->dimensions[1]) {%(fail)s} for (size_t i = 0; i < Nx[0]; ++i) { size_t j; double sum = 0.0; bool discount_max = false; const double* __restrict__ x_i = (double*)(%(x)s->data + %(x)s->strides[0] * i); const double* __restrict__ b_i = (double*)(%(b)s->data); const long int y_i = ((long int*)(%(y_idx)s->data + %(y_idx)s->strides[0] * i))[0]; double* __restrict__ sm_i = (double*)(%(sm)s->data + %(sm)s->strides[0] * i); double* __restrict__ nll_i = (double*)(%(nll)s->data + %(nll)s->strides[0] * i); npy_intp Sx = %(x)s->strides[1]/sizeof(double); npy_intp Sb = %(b)s->strides[0]/sizeof(double); npy_intp Ssm = %(sm)s->strides[1]/sizeof(double); size_t row_max_j=0; double row_max = x_i[0] + b_i[0]; //try to compute sum and sm the easy way for (j = 0; j < Nx[1]; ++j) { double row_ij = x_i[j * Sx] + b_i[j * Sb]; row_max_j = (row_ij > row_max) ? j : row_max_j; row_max = (row_ij > row_max) ? row_ij : row_max; double sm_ij = exp(row_ij); sum += sm_ij; sm_i[j * Ssm] = sm_ij; } if ((0.0 == sum) || (isinf(sum))) { //our cheap trick didn't work... try again and do it better. discount_max = true; sum = 0.0; //reset sum and recompute.... for (j = 0; j < Nx[1]; ++j) { double row_ij = x_i[j * Sx] + b_i[j * Sb]; double sm_ij = exp(row_ij - row_max); sum += sm_ij; sm_i[j * Ssm] = sm_ij; } if ( (0.0 == sum) || (isinf(sum))) { //that was our best... %(fail)s; } //if we still can't sum it up, we're screwed. //So far, this assertion has never failed... } //cblas_dscal(x.N, 1.0 / sum, &mat_at(s,i,0), s.n); double sum_inv = 1.0 / sum; for (j = 0; j < Nx[1]; ++j) { sm_i[j * Ssm] *= sum_inv; } if (y_i >= Nx[1]) { %(fail)s; } nll_i[0] = - x_i[y_i*Sx] - b_i[y_i*Sb] + (discount_max ? row_max : 0.0) + log(sum); //mat_at(y,i,0) = -log( mat_at(s,i,t[i])); //less accurate? //mat_at(y,i,0) = - mat_at(x,i,t[i]) - mat_at(b,0,t[i]) + (discount_max ? maxi : 0.0) + log(sum); } """ % dict(locals(), **sub) crossentropy_softmax_1hot = gof.op.constructor(CrossentropySoftmax1Hot) class CrossentropySoftmax1HotDx (gof.op.Op): nin=3 nout=1 """Gradient wrt x of the CrossentropySoftmax1Hot Op""" def __init__(self, dy, sm, y_idx,**kwargs): dy = tensor._as_tensor(dy) sm = tensor._as_tensor(sm) y_idx = tensor._as_tensor(y_idx) self.inputs = [dy, sm, y_idx] self.outputs = [tensor.Tensor(sm.dtype, sm.broadcastable)] def perform(self): dy,sm,y_idx = [i.data for i in self.inputs] dx = numpy.zeros_like(sm) for i in xrange(sm.shape[0]): dx[i] = dy[i] * sm[i] #vector scale dx[i, y_idx[i]] -= dy[i] #scalar decrement self.outputs[0].data = dx def grad(self, *args): raise NotImplementedError() def c_validate_update(self, (dnll, sm, y_idx), (dx,), sub): """Allocate output storage""" return """ if (%(dnll)s->nd != 1) { %(fail)s } if (%(sm)s->nd != 2) { %(fail)s } if (%(y_idx)s->nd != 1) { %(fail)s } if (%(dnll)s->descr->type_num != PyArray_DOUBLE) { %(fail)s} if (%(sm)s->descr->type_num != PyArray_DOUBLE) { %(fail)s} if (%(y_idx)s->descr->type_num != PyArray_INT64) { %(fail)s} %(dx)s = (PyArrayObject*)PyArray_SimpleNew(2, PyArray_DIMS(%(sm)s), type_num_%(sm)s); if(!%(dx)s){%(fail)s} """ % dict(locals(), **sub) def c_validate_cleanup(self, inputs, outputs, sub): """Not sure...""" return "" def c_support_code(self): return """ """ def c_code(self, (dnll, sm, y_idx), (dx,), sub): return """ npy_intp* shape = %(dx)s->dimensions; if (%(dnll)s->dimensions[0] != %(sm)s->dimensions[0]) {%(fail)s} if (%(dnll)s->dimensions[0] != %(y_idx)s->dimensions[0]) {%(fail)s} if (%(dnll)s->dimensions[0] != %(dx)s->dimensions[0]) {%(fail)s} if (%(sm)s->dimensions[1] != %(dx)s->dimensions[1]) {%(fail)s} for (size_t i = 0; i < shape[0]; ++i) { const double dnll_i = ((double*)(%(dnll)s->data + %(dnll)s->strides[0] * i))[0]; const long int y_i = ((long int*)(%(y_idx)s->data + %(y_idx)s->strides[0] * i))[0]; const double* __restrict__ sm_i = (double*)(%(sm)s->data + %(sm)s->strides[0] * i); npy_intp Ssm = %(sm)s->strides[1]/sizeof(double); double* __restrict__ dx_i = (double*)(%(dx)s->data + %(dx)s->strides[0] * i); npy_intp Sdx = %(dx)s->strides[1]/sizeof(double); for (size_t j = 0; j < shape[1]; ++j) { dx_i[j * Sdx] = dnll_i * sm_i[j * Ssm]; } if (y_i >= shape[1]) { %(fail)s; } dx_i[y_i * Sdx] -= dnll_i; } """ % dict(locals(), **sub)