Loading deepchem/rl/__init__.py +20 −1 Original line number Diff line number Diff line Loading @@ -19,12 +19,22 @@ class Environment(object): running in different processes or even on different computers. """ def __init__(self, state_shape, n_actions): def __init__(self, state_shape, n_actions, state_dtype=None): """Subclasses should call the superclass constructor in addition to doing their own initialization.""" self._state_shape = state_shape self._n_actions = n_actions self._state = None self._terminated = None if state_dtype is None: # Assume all arrays are float32. import numpy import collections if isinstance(state_shape[0], collections.Sequence): self._state_dtype = [numpy.float32] * len(state_shape) else: self._state_dtype = numpy.float32 else: self._state_dtype = state_dtype @property def state(self): Loading Loading @@ -52,6 +62,15 @@ class Environment(object): """ return self._state_shape @property def state_dtype(self): """The dtypes of the arrays that describe a state. If the state is a single array, this returns the dtype of that array. If the state is a list of arrays, this returns a list containing the dtypes of the arrays. """ return self._state_dtype @property def n_actions(self): """The number of possible actions that can be performed in this Environment.""" Loading deepchem/rl/a3c.py +5 −1 Original line number Diff line number Diff line Loading @@ -138,9 +138,13 @@ class A3C(object): def _build_graph(self, tf_graph, scope, model_dir): """Construct a TensorGraph containing the policy and loss calculations.""" state_shape = self._env.state_shape state_dtype = self._env.state_dtype if not self._state_is_list: state_shape = [state_shape] features = [Feature(shape=[None] + list(s)) for s in state_shape] state_dtype = [state_dtype] features = [] for s, d in zip(state_shape, state_dtype): features.append(Feature(shape=[None] + list(s), dtype=tf.as_dtype(d))) policy_layers = self._policy.create_layers(features) action_prob = policy_layers['action_prob'] value = policy_layers['value'] Loading deepchem/rl/ppo.py +5 −1 Original line number Diff line number Diff line Loading @@ -160,9 +160,13 @@ class PPO(object): def _build_graph(self, tf_graph, scope, model_dir): """Construct a TensorGraph containing the policy and loss calculations.""" state_shape = self._env.state_shape state_dtype = self._env.state_dtype if not self._state_is_list: state_shape = [state_shape] features = [Feature(shape=[None] + list(s)) for s in state_shape] state_dtype = [state_dtype] features = [] for s, d in zip(state_shape, state_dtype): features.append(Feature(shape=[None] + list(s), dtype=tf.as_dtype(d))) policy_layers = self._policy.create_layers(features) action_prob = policy_layers['action_prob'] value = policy_layers['value'] Loading Loading
deepchem/rl/__init__.py +20 −1 Original line number Diff line number Diff line Loading @@ -19,12 +19,22 @@ class Environment(object): running in different processes or even on different computers. """ def __init__(self, state_shape, n_actions): def __init__(self, state_shape, n_actions, state_dtype=None): """Subclasses should call the superclass constructor in addition to doing their own initialization.""" self._state_shape = state_shape self._n_actions = n_actions self._state = None self._terminated = None if state_dtype is None: # Assume all arrays are float32. import numpy import collections if isinstance(state_shape[0], collections.Sequence): self._state_dtype = [numpy.float32] * len(state_shape) else: self._state_dtype = numpy.float32 else: self._state_dtype = state_dtype @property def state(self): Loading Loading @@ -52,6 +62,15 @@ class Environment(object): """ return self._state_shape @property def state_dtype(self): """The dtypes of the arrays that describe a state. If the state is a single array, this returns the dtype of that array. If the state is a list of arrays, this returns a list containing the dtypes of the arrays. """ return self._state_dtype @property def n_actions(self): """The number of possible actions that can be performed in this Environment.""" Loading
deepchem/rl/a3c.py +5 −1 Original line number Diff line number Diff line Loading @@ -138,9 +138,13 @@ class A3C(object): def _build_graph(self, tf_graph, scope, model_dir): """Construct a TensorGraph containing the policy and loss calculations.""" state_shape = self._env.state_shape state_dtype = self._env.state_dtype if not self._state_is_list: state_shape = [state_shape] features = [Feature(shape=[None] + list(s)) for s in state_shape] state_dtype = [state_dtype] features = [] for s, d in zip(state_shape, state_dtype): features.append(Feature(shape=[None] + list(s), dtype=tf.as_dtype(d))) policy_layers = self._policy.create_layers(features) action_prob = policy_layers['action_prob'] value = policy_layers['value'] Loading
deepchem/rl/ppo.py +5 −1 Original line number Diff line number Diff line Loading @@ -160,9 +160,13 @@ class PPO(object): def _build_graph(self, tf_graph, scope, model_dir): """Construct a TensorGraph containing the policy and loss calculations.""" state_shape = self._env.state_shape state_dtype = self._env.state_dtype if not self._state_is_list: state_shape = [state_shape] features = [Feature(shape=[None] + list(s)) for s in state_shape] state_dtype = [state_dtype] features = [] for s, d in zip(state_shape, state_dtype): features.append(Feature(shape=[None] + list(s), dtype=tf.as_dtype(d))) policy_layers = self._policy.create_layers(features) action_prob = policy_layers['action_prob'] value = policy_layers['value'] Loading
