Python ndarray.slice_axis方法代码示例

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def argtopk(input, k, dim, descending=True):
    idx = nd.argsort(input, dim, is_ascend=not descending)
    return nd.slice_axis(input, dim, 0, k) 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def slice_axis(data, axis, begin, end):
    dim = data.shape[axis]
    if begin < 0:
        begin += dim
    if end <= 0:
        end += dim
    return nd.slice_axis(data, axis, begin, end) 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def dumpR(data_set, mx_model, batch_size, name='', data_extra = None, label_shape = None):
  print('dump verification embedding..')
  data_list = data_set[0]
  issame_list = data_set[1]
  model = mx_model
  embeddings_list = []
  if data_extra is not None:
    _data_extra = nd.array(data_extra)
  time_consumed = 0.0
  if label_shape is None:
    _label = nd.ones( (batch_size,) )
  else:
    _label = nd.ones( label_shape )
  for i in range( len(data_list) ):
    data = data_list[i]
    embeddings = None
    ba = 0
    while ba<data.shape[0]:
      bb = min(ba+batch_size, data.shape[0])
      count = bb-ba
      _data = nd.slice_axis(data, axis=0, begin=bb-batch_size, end=bb)
      #print(_data.shape, _label.shape)
      time0 = datetime.datetime.now()
      if data_extra is None:
        db = mx.io.DataBatch(data=(_data,), label=(_label,))
      else:
        db = mx.io.DataBatch(data=(_data,_data_extra), label=(_label,))
      model.forward(db, is_train=False)
      net_out = model.get_outputs()
      _embeddings = net_out[0].asnumpy()
      time_now = datetime.datetime.now()
      diff = time_now - time0
      time_consumed+=diff.total_seconds()
      if embeddings is None:
        embeddings = np.zeros( (data.shape[0], _embeddings.shape[1]) )
      embeddings[ba:bb,:] = _embeddings[(batch_size-count):,:]
      ba = bb
    embeddings_list.append(embeddings)
  embeddings = embeddings_list[0] + embeddings_list[1]
  embeddings = sklearn.preprocessing.normalize(embeddings)
  actual_issame = np.asarray(issame_list)
  outname = os.path.join('temp.bin')
  with open(outname, 'wb') as f:
    pickle.dump((embeddings, issame_list), f, protocol=pickle.HIGHEST_PROTOCOL) 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def test(lfw_set, mx_model, batch_size):
  print('testing lfw..')
  lfw_data_list = lfw_set[0]
  issame_list = lfw_set[1]
  model = mx_model
  embeddings_list = []
  for i in range( len(lfw_data_list) ):
    lfw_data = lfw_data_list[i]
    embeddings = None
    ba = 0
    while ba<lfw_data.shape[0]:
      bb = min(ba+batch_size, lfw_data.shape[0])
      _data = nd.slice_axis(lfw_data, axis=0, begin=ba, end=bb)
      _label = nd.ones( (bb-ba,) )
      #print(_data.shape, _label.shape)
      db = mx.io.DataBatch(data=(_data,), label=(_label,))
      model.forward(db, is_train=False)
      net_out = model.get_outputs()
      #_arg, _aux = model.get_params()
      #__arg = {}
      #for k,v in _arg.iteritems():
      #  __arg[k] = v.as_in_context(_ctx)
      #_arg = __arg
      #_arg["data"] = _data.as_in_context(_ctx)
      #_arg["softmax_label"] = _label.as_in_context(_ctx)
      #for k,v in _arg.iteritems():
      #  print(k,v.context)
      #exe = sym.bind(_ctx, _arg ,args_grad=None, grad_req="null", aux_states=_aux)
      #exe.forward(is_train=False)
      #net_out = exe.outputs
      _embeddings = net_out[0].asnumpy()
      #print(_embeddings.shape)
      if embeddings is None:
        embeddings = np.zeros( (lfw_data.shape[0], _embeddings.shape[1]) )
      embeddings[ba:bb,:] = _embeddings
      ba = bb
    embeddings_list.append(embeddings)

  _xnorm = 0.0
  _xnorm_cnt = 0
  for embed in embeddings_list:
    for i in range(embed.shape[0]):
      _em = embed[i]
      _norm=np.linalg.norm(_em)
      #print(_em.shape, _norm)
      _xnorm+=_norm
      _xnorm_cnt+=1
  _xnorm /= _xnorm_cnt

  embeddings = embeddings_list[0].copy()
  embeddings = sklearn.preprocessing.normalize(embeddings)
  _, _, accuracy, val, val_std, far = evaluate(embeddings, issame_list, nrof_folds=10)
  acc1, std1 = np.mean(accuracy), np.std(accuracy)
  #print('Validation rate: %2.5f+-%2.5f @ FAR=%2.5f' % (val, val_std, far))
  #embeddings = np.concatenate(embeddings_list, axis=1)
  embeddings = embeddings_list[0] + embeddings_list[1]
  embeddings = sklearn.preprocessing.normalize(embeddings)
  print(embeddings.shape)
  _, _, accuracy, val, val_std, far = evaluate(embeddings, issame_list, nrof_folds=10)
  acc2, std2 = np.mean(accuracy), np.std(accuracy)
  return acc1, std1, acc2, std2, _xnorm, embeddings_list 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def dumpR(data_set, mx_model, batch_size, name='', data_extra = None, label_shape = None):
  print('dump verification embedding..')
  data_list = data_set[0]
  issame_list = data_set[1]
  model = mx_model
  embeddings_list = []
  if data_extra is not None:
    _data_extra = nd.array(data_extra)
  time_consumed = 0.0
  if label_shape is None:
    _label = nd.ones( (batch_size,) )
  else:
    _label = nd.ones( label_shape )
  for i in xrange( len(data_list) ):
    data = data_list[i]
    embeddings = None
    ba = 0
    while ba<data.shape[0]:
      bb = min(ba+batch_size, data.shape[0])
      count = bb-ba
      _data = nd.slice_axis(data, axis=0, begin=bb-batch_size, end=bb)
      #print(_data.shape, _label.shape)
      time0 = datetime.datetime.now()
      if data_extra is None:
        db = mx.io.DataBatch(data=(_data,), label=(_label,))
      else:
        db = mx.io.DataBatch(data=(_data,_data_extra), label=(_label,))
      model.forward(db, is_train=False)
      net_out = model.get_outputs()
      _embeddings = net_out[0].asnumpy()
      time_now = datetime.datetime.now()
      diff = time_now - time0
      time_consumed+=diff.total_seconds()
      if embeddings is None:
        embeddings = np.zeros( (data.shape[0], _embeddings.shape[1]) )
      embeddings[ba:bb,:] = _embeddings[(batch_size-count):,:]
      ba = bb
    embeddings_list.append(embeddings)
  embeddings = embeddings_list[0] + embeddings_list[1]
  embeddings = sklearn.preprocessing.normalize(embeddings)
  actual_issame = np.asarray(issame_list)
  outname = os.path.join('temp.bin')
  with open(outname, 'wb') as f:
    pickle.dump((embeddings, issame_list), f, protocol=pickle.HIGHEST_PROTOCOL) 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def test(lfw_set, mx_model, batch_size):
  print('testing lfw..')
  lfw_data_list = lfw_set[0]
  issame_list = lfw_set[1]
  model = mx_model
  embeddings_list = []
  for i in xrange( len(lfw_data_list) ):
    lfw_data = lfw_data_list[i]
    embeddings = None
    ba = 0
    while ba<lfw_data.shape[0]:
      bb = min(ba+batch_size, lfw_data.shape[0])
      _data = nd.slice_axis(lfw_data, axis=0, begin=ba, end=bb)
      _label = nd.ones( (bb-ba,) )
      #print(_data.shape, _label.shape)
      db = mx.io.DataBatch(data=(_data,), label=(_label,))
      model.forward(db, is_train=False)
      net_out = model.get_outputs()
      #_arg, _aux = model.get_params()
      #__arg = {}
      #for k,v in _arg.iteritems():
      #  __arg[k] = v.as_in_context(_ctx)
      #_arg = __arg
      #_arg["data"] = _data.as_in_context(_ctx)
      #_arg["softmax_label"] = _label.as_in_context(_ctx)
      #for k,v in _arg.iteritems():
      #  print(k,v.context)
      #exe = sym.bind(_ctx, _arg ,args_grad=None, grad_req="null", aux_states=_aux)
      #exe.forward(is_train=False)
      #net_out = exe.outputs
      _embeddings = net_out[0].asnumpy()
      #print(_embeddings.shape)
      if embeddings is None:
        embeddings = np.zeros( (lfw_data.shape[0], _embeddings.shape[1]) )
      embeddings[ba:bb,:] = _embeddings
      ba = bb
    embeddings_list.append(embeddings)

  _xnorm = 0.0
  _xnorm_cnt = 0
  for embed in embeddings_list:
    for i in xrange(embed.shape[0]):
      _em = embed[i]
      _norm=np.linalg.norm(_em)
      #print(_em.shape, _norm)
      _xnorm+=_norm
      _xnorm_cnt+=1
  _xnorm /= _xnorm_cnt

  embeddings = embeddings_list[0].copy()
  embeddings = sklearn.preprocessing.normalize(embeddings)
  _, _, accuracy, val, val_std, far = evaluate(embeddings, issame_list, nrof_folds=10)
  acc1, std1 = np.mean(accuracy), np.std(accuracy)
  #print('Validation rate: %2.5f+-%2.5f @ FAR=%2.5f' % (val, val_std, far))
  #embeddings = np.concatenate(embeddings_list, axis=1)
  embeddings = embeddings_list[0] + embeddings_list[1]
  embeddings = sklearn.preprocessing.normalize(embeddings)
  print(embeddings.shape)
  _, _, accuracy, val, val_std, far = evaluate(embeddings, issame_list, nrof_folds=10)
  acc2, std2 = np.mean(accuracy), np.std(accuracy)
  return acc1, std1, acc2, std2, _xnorm, embeddings_list 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def split_data(data, num_slice, batch_axis=0, even_split=True, multiplier=1):
    """Splits an NDArray into `num_slice` slices along `batch_axis`.
    Usually used for data parallelism where each slices is sent
    to one device (i.e. GPU).

    Parameters
    ----------
    data : NDArray
        A batch of data.
    num_slice : int
        Number of desired slices.
    batch_axis : int, default 0
        The axis along which to slice.
    even_split : bool, default True
        Whether to force all slices to have the same number of elements.
        If `True`, an error will be raised when `num_slice` does not evenly
        divide `data.shape[batch_axis]`.
    multiplier : int, default 1
        The batch size has to be the multiples of multiplier

    Returns
    -------
    list of NDArray
        Return value is a list even if `num_slice` is 1.
    """
    size = data.shape[batch_axis]
    if even_split and size % num_slice != 0:
        raise ValueError(
            "data with shape %s cannot be evenly split into %d slices along axis %d. " \
            "Use a batch size that's multiple of %d or set even_split=False to allow " \
            "uneven partitioning of data."%(
                str(data.shape), num_slice, batch_axis, num_slice))

    step = (int(size / multiplier) // num_slice) * multiplier

    # If size < num_slice, make fewer slices
    if not even_split and size < num_slice:
        step = 1
        num_slice = size

    if batch_axis == 0:
        slices = [data[i*step:(i+1)*step] if i < num_slice - 1 else data[i*step:size]
                  for i in range(num_slice)]
    elif even_split:
        slices = ndarray.split(data, num_outputs=num_slice, axis=batch_axis)
    else:
        slices = [ndarray.slice_axis(data, batch_axis, i*step, (i+1)*step)
                  if i < num_slice - 1 else
                  ndarray.slice_axis(data, batch_axis, i*step, size)
                  for i in range(num_slice)]
    return slices 

# 需要导入模块: from mxnet import ndarray [as 别名]
# 或者: from mxnet.ndarray import slice_axis [as 别名]
def dumpR(data_set, mx_model, batch_size, name='', data_extra=None, label_shape=None):
    print('dump verification embedding..')
    data_list = data_set[0]
    issame_list = data_set[1]
    model = mx_model
    embeddings_list = []
    if data_extra is not None:
        _data_extra = nd.array(data_extra)
    time_consumed = 0.0
    if label_shape is None:
        _label = nd.ones((batch_size,))
    else:
        _label = nd.ones(label_shape)
    for i in range(len(data_list)):
        data = data_list[i]
        embeddings = None
        ba = 0
        while ba < data.shape[0]:
            bb = min(ba + batch_size, data.shape[0])
            count = bb - ba
            _data = nd.slice_axis(data, axis=0, begin=bb - batch_size, end=bb)
            # print(_data.shape, _label.shape)
            time0 = datetime.datetime.now()
            if data_extra is None:
                db = mx.io.DataBatch(data=(_data,), label=(_label,))
            else:
                db = mx.io.DataBatch(data=(_data, _data_extra), label=(_label,))
            model.forward(db, is_train=False)
            net_out = model.get_outputs()
            _embeddings = net_out[0].asnumpy()
            time_now = datetime.datetime.now()
            diff = time_now - time0
            time_consumed += diff.total_seconds()
            if embeddings is None:
                embeddings = np.zeros((data.shape[0], _embeddings.shape[1]))
            embeddings[ba:bb, :] = _embeddings[(batch_size - count):, :]
            ba = bb
        embeddings_list.append(embeddings)
    embeddings = embeddings_list[0] + embeddings_list[1]
    embeddings = sklearn.preprocessing.normalize(embeddings)
    actual_issame = np.asarray(issame_list)
    outname = os.path.join('temp.bin')
    with open(outname, 'wb') as f:
        pickle.dump((embeddings, issame_list), f, protocol=pickle.HIGHEST_PROTOCOL)