Commit 13d859ec authored by miaecle's avatar miaecle
Browse files

recover breaking changes

parent 83ddf864

deepchem/models/tf_new_models/multitask_classifier.py

+6 −8
Original line number Diff line number Diff line
@@ -30,23 +30,23 @@ def get_loss_fn(final_loss): 
  if final_loss == 'L2':



    def loss_fn(x, t):

      diff = tf.substract(x, t)

      diff = tf.subtract(x, t)

      return tf.reduce_sum(tf.square(diff), 0)

  elif final_loss == 'weighted_L2':



    def loss_fn(x, t, w):

      diff = tf.substract(x, t)

      diff = tf.subtract(x, t)

      weighted_diff = tf.multiply(diff, w)

      return tf.reduce_sum(tf.square(weighted_diff), 0)

  elif final_loss == 'L1':



    def loss_fn(x, t):

      diff = tf.substract(x, t)

      diff = tf.subtract(x, t)

      return tf.reduce_sum(tf.abs(diff), 0)

  elif final_loss == 'huber':



    def loss_fn(x, t):

      diff = tf.substract(x, t)

      diff = tf.subtract(x, t)

      return tf.reduce_sum(

          tf.minimum(0.5 * tf.square(diff),

                     huber_d * (tf.abs(diff) - 0.5 * huber_d)), 0)
@@ -188,10 +188,8 @@ class MultitaskGraphClassifier(Model): 
    task_losses = []

    # label_placeholder of shape (batch_size, n_tasks). Split into n_tasks

    # tensors of shape (batch_size,)

    task_labels = tf.split(

        axis=1, num_or_size_splits=self.n_tasks, value=self.label_placeholder)

    task_weights = tf.split(

        axis=1, num_or_size_splits=self.n_tasks, value=self.weight_placeholder)

    task_labels = tf.split(1, self.n_tasks, self.label_placeholder)

    task_weights = tf.split(1, self.n_tasks, self.weight_placeholder)

    for task in range(self.n_tasks):

      task_label_vector = task_labels[task]

      task_weight_vector = task_weights[task]

deepchem/models/tf_new_models/multitask_regressor.py

+2 −4
Original line number Diff line number Diff line
@@ -137,10 +137,8 @@ class MultitaskGraphRegressor(Model): 
    task_losses = []

    # label_placeholder of shape (batch_size, n_tasks). Split into n_tasks

    # tensors of shape (batch_size,)

    task_labels = tf.split(

        axis=1, num_or_size_splits=self.n_tasks, value=self.label_placeholder)

    task_weights = tf.split(

        axis=1, num_or_size_splits=self.n_tasks, value=self.weight_placeholder)

    task_labels = tf.split(1, self.n_tasks, self.label_placeholder)

    task_weights = tf.split(1, self.n_tasks, self.weight_placeholder)

    for task in range(self.n_tasks):

      task_label_vector = task_labels[task]

      task_weight_vector = task_weights[task]

deepchem/nn/layers.py

+11 −11
Original line number Diff line number Diff line
@@ -85,8 +85,8 @@ def graph_conv(atoms, deg_adj_lists, deg_slice, max_deg, min_deg, W_list, 
    rel_atoms = deg_summed[deg - 1]



    # Get self atoms

    begin = tf.stack([deg_slice[deg - min_deg, 0], 0])

    size = tf.stack([deg_slice[deg - min_deg, 1], -1])

    begin = tf.pack([deg_slice[deg - min_deg, 0], 0])

    size = tf.pack([deg_slice[deg - min_deg, 1], -1])

    self_atoms = tf.slice(atoms, begin, size)



    # Apply hidden affine to relevant atoms and append
@@ -110,7 +110,7 @@ def graph_conv(atoms, deg_adj_lists, deg_slice, max_deg, min_deg, W_list, 
    new_rel_atoms_collection[deg - min_deg] = out



  # Combine all atoms back into the list

  activated_atoms = tf.concat(axis=0, values=new_rel_atoms_collection)

  activated_atoms = tf.concat(0, new_rel_atoms_collection)



  return activated_atoms
@@ -145,7 +145,7 @@ def graph_gather(atoms, membership_placeholder, batch_size): 
  ]



  # Get the final sparse representations

  sparse_reps = tf.concat(axis=0, values=sparse_reps)

  sparse_reps = tf.concat(0, sparse_reps)



  return sparse_reps
@@ -178,8 +178,8 @@ def graph_pool(atoms, deg_adj_lists, deg_slice, max_deg, min_deg): 


  for deg in range(1, max_deg + 1):

    # Get self atoms

    begin = tf.stack([deg_slice[deg - min_deg, 0], 0])

    size = tf.stack([deg_slice[deg - min_deg, 1], -1])

    begin = tf.pack([deg_slice[deg - min_deg, 0], 0])

    size = tf.pack([deg_slice[deg - min_deg, 1], -1])

    self_atoms = tf.slice(atoms, begin, size)



    # Expand dims
@@ -187,18 +187,18 @@ def graph_pool(atoms, deg_adj_lists, deg_slice, max_deg, min_deg): 


    # always deg-1 for deg_adj_lists

    gathered_atoms = tf.gather(atoms, deg_adj_lists[deg - 1])

    gathered_atoms = tf.concat(axis=1, values=[self_atoms, gathered_atoms])

    gathered_atoms = tf.concat(1, [self_atoms, gathered_atoms])



    maxed_atoms = tf.reduce_max(gathered_atoms, 1)

    deg_maxed[deg - min_deg] = maxed_atoms



  if min_deg == 0:

    begin = tf.stack([deg_slice[0, 0], 0])

    size = tf.stack([deg_slice[0, 1], -1])

    begin = tf.pack([deg_slice[0, 0], 0])

    size = tf.pack([deg_slice[0, 1], -1])

    self_atoms = tf.slice(atoms, begin, size)

    deg_maxed[0] = self_atoms



  return tf.concat(axis=0, values=deg_maxed)

  return tf.concat(0, deg_maxed)





class GraphConv(Layer):

deepchem/nn/model_ops.py

+10 −10
Original line number Diff line number Diff line
@@ -218,7 +218,7 @@ def concatenate(tensors, axis=-1): 
  try:

    return tf.concat_v2([x for x in tensors], axis)

  except AttributeError:

    return tf.concat(axis=axis, values=[x for x in tensors])

    return tf.concat(axis, [x for x in tensors])





def _normalize_axis(axis, ndim):
@@ -253,7 +253,7 @@ def mean(x, axis=None, keepdims=False): 
  axis = _normalize_axis(axis, get_ndim(x))

  if x.dtype.base_dtype == tf.bool:

    x = tf.cast(x, tf.float32)

  return tf.reduce_mean(x, axis=axis, keep_dims=keepdims)

  return tf.reduce_mean(x, reduction_indices=axis, keep_dims=keepdims)





def dot(x, y):
@@ -273,14 +273,14 @@ def dot(x, y): 
  """

  if get_ndim(x) is not None and (get_ndim(x) > 2 or get_ndim(y) > 2):

    x_shape = []

    for i, s in zip(int_shape(x), tf.unstack(tf.shape(x))):

    for i, s in zip(int_shape(x), tf.unpack(tf.shape(x))):

      if i is not None:

        x_shape.append(i)

      else:

        x_shape.append(s)

    x_shape = tuple(x_shape)

    y_shape = []

    for i, s in zip(int_shape(y), tf.unstack(tf.shape(y))):

    for i, s in zip(int_shape(y), tf.unpack(tf.shape(y))):

      if i is not None:

        y_shape.append(i)

      else:
@@ -431,7 +431,7 @@ def max(x, axis=None, keepdims=False): 
  A tensor with maximum values of `x`.

  """

  axis = _normalize_axis(axis, get_ndim(x))

  return tf.reduce_max(x, axis=axis, keep_dims=keepdims)

  return tf.reduce_max(x, reduction_indices=axis, keep_dims=keepdims)





def l2_normalize(x, axis):
@@ -463,12 +463,12 @@ def categorical_crossentropy(output, target, from_logits=False): 
  if not from_logits:

    # scale preds so that the class probas of each sample sum to 1

    output /= tf.reduce_sum(

        output, axis=len(output.get_shape()) - 1, keep_dims=True)

        output, reduction_indices=len(output.get_shape()) - 1, keep_dims=True)

    # manual computation of crossentropy

    epsilon = _to_tensor(_EPSILON, output.dtype.base_dtype)

    output = tf.clip_by_value(output, epsilon, 1. - epsilon)

    return -tf.reduce_sum(

        target * tf.log(output), axis=len(output.get_shape()) - 1)

        target * tf.log(output), reduction_indices=len(output.get_shape()) - 1)

  else:

    try:

      return tf.nn.softmax_cross_entropy_with_logits(
@@ -700,10 +700,10 @@ def var(x, axis=None, keepdims=False): 
  axis = _normalize_axis(axis, get_ndim(x))

  if x.dtype.base_dtype == tf.bool:

    x = tf.cast(x, tf.float32)

  m = tf.reduce_mean(x, axis=axis, keep_dims=True)

  m = tf.reduce_mean(x, reduction_indices=axis, keep_dims=True)

  devs_squared = tf.square(x - m)

  return tf.reduce_mean(

      devs_squared, axis=axis, keep_dims=keepdims)

      devs_squared, reduction_indices=axis, keep_dims=keepdims)





def euclidean_distance(test, support, max_dist_sq=20):
@@ -959,7 +959,7 @@ def softmax_N(tensor, name=None): 
    return tf.div(exp_tensor,

                  tf.reduce_sum(

                      exp_tensor,

                      axis=reduction_indices,

                      reduction_indices=reduction_indices,

                      keep_dims=True))

CRA Git | Maintained and supported by SUSTech CRA and CCSE