Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

polish(pu): polish reward/value/policy_head_hidden_channels #314

Merged
merged 1 commit into from
Jan 8, 2025
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions lzero/agent/config/alphazero/tictactoe_play_with_bot.py
Original file line number Diff line number Diff line change
Expand Up @@ -53,8 +53,8 @@
# We use the small size model for tictactoe.
num_res_blocks=1,
num_channels=16,
fc_value_layers=[8],
fc_policy_layers=[8],
value_head_hidden_channels=[8],
policy_head_hidden_channels=[8],
),
cuda=True,
board_size=3,
Expand Down
6 changes: 3 additions & 3 deletions lzero/agent/config/gumbel_muzero/tictactoe_play_with_bot.py
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,9 @@
# We use the small size model for tictactoe.
num_res_blocks=1,
num_channels=16,
fc_reward_layers=[8],
fc_value_layers=[8],
fc_policy_layers=[8],
reward_head_hidden_channels=[8],
value_head_hidden_channels=[8],
policy_head_hidden_channels=[8],
support_scale=10,
reward_support_size=21,
value_support_size=21,
Expand Down
6 changes: 3 additions & 3 deletions lzero/agent/config/muzero/tictactoe_play_with_bot.py
Original file line number Diff line number Diff line change
Expand Up @@ -35,9 +35,9 @@
# We use the small size model for tictactoe.
num_res_blocks=1,
num_channels=16,
fc_reward_layers=[8],
fc_value_layers=[8],
fc_policy_layers=[8],
reward_head_hidden_channels=[8],
value_head_hidden_channels=[8],
policy_head_hidden_channels=[8],
support_scale=10,
reward_support_size=21,
value_support_size=21,
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -52,8 +52,8 @@
# We use the small size model for tictactoe.
num_res_blocks=1,
num_channels=16,
fc_value_layers=[8],
fc_policy_layers=[8],
value_head_hidden_channels=[8],
policy_head_hidden_channels=[8],
),
sampled_algo=True,
mcts_ctree=mcts_ctree,
Expand Down
Original file line number Diff line number Diff line change
Expand Up @@ -50,9 +50,9 @@
num_channels=16,
frame_stack_num=1,
model_type='conv',
fc_reward_layers=[8],
fc_value_layers=[8],
fc_policy_layers=[8],
reward_head_hidden_channels=[8],
value_head_hidden_channels=[8],
policy_head_hidden_channels=[8],
support_scale=10,
reward_support_size=21,
value_support_size=21,
Expand Down
64 changes: 32 additions & 32 deletions lzero/model/alphazero_model.py
Original file line number Diff line number Diff line change
Expand Up @@ -32,8 +32,8 @@ def __init__(
num_channels: int = 64,
value_head_channels: int = 16,
policy_head_channels: int = 16,
fc_value_layers: SequenceType = [32],
fc_policy_layers: SequenceType = [32],
value_head_hidden_channels: SequenceType = [32],
policy_head_hidden_channels: SequenceType = [32],
value_support_size: int = 601,
# ==============================================================
# specific sampled related config
Expand Down Expand Up @@ -66,8 +66,8 @@ def __init__(
- num_channels (:obj:`int`): The channels of hidden states.
- value_head_channels (:obj:`int`): The channels of value head.
- policy_head_channels (:obj:`int`): The channels of policy head.
- fc_value_layers (:obj:`SequenceType`): The number of hidden layers used in value head (MLP head).
- fc_policy_layers (:obj:`SequenceType`): The number of hidden layers used in policy head (MLP head).
- value_head_hidden_channels (:obj:`SequenceType`): The number of hidden layers used in value head (MLP head).
- policy_head_hidden_channels (:obj:`SequenceType`): The number of hidden layers used in policy head (MLP head).
- value_support_size (:obj:`int`): The size of categorical value.
"""
super(AlphaZeroModel, self).__init__()
Expand Down Expand Up @@ -102,14 +102,14 @@ def __init__(
self.num_of_sampled_actions = num_of_sampled_actions

# TODO use more adaptive way to get the flatten output size
flatten_output_size_for_value_head = (
flatten_input_size_for_value_head = (
(
value_head_channels * math.ceil(self.observation_shape[1] / 16) *
math.ceil(self.observation_shape[2] / 16)
) if downsample else (value_head_channels * self.observation_shape[1] * self.observation_shape[2])
)

flatten_output_size_for_policy_head = (
flatten_input_size_for_policy_head = (
(
policy_head_channels * math.ceil(self.observation_shape[1] / 16) *
math.ceil(self.observation_shape[2] / 16)
Expand All @@ -123,11 +123,11 @@ def __init__(
num_channels,
value_head_channels,
policy_head_channels,
fc_value_layers,
fc_policy_layers,
value_head_hidden_channels,
policy_head_hidden_channels,
self.value_support_size,
flatten_output_size_for_value_head,
flatten_output_size_for_policy_head,
flatten_input_size_for_value_head,
flatten_input_size_for_policy_head,
last_linear_layer_init_zero=self.last_linear_layer_init_zero,
activation=activation,
sigma_type=self.sigma_type,
Expand Down Expand Up @@ -216,11 +216,11 @@ def __init__(
num_channels: int,
value_head_channels: int,
policy_head_channels: int,
fc_value_layers: SequenceType,
fc_policy_layers: SequenceType,
value_head_hidden_channels: SequenceType,
policy_head_hidden_channels: SequenceType,
output_support_size: int,
flatten_output_size_for_value_head: int,
flatten_output_size_for_policy_head: int,
flatten_input_size_for_value_head: int,
flatten_input_size_for_policy_head: int,
last_linear_layer_init_zero: bool = True,
activation: Optional[nn.Module] = nn.ReLU(inplace=True),
# ==============================================================
Expand All @@ -241,12 +241,12 @@ def __init__(
- num_channels (:obj:`int`): The channels of hidden states.
- value_head_channels (:obj:`int`): The channels of value head.
- policy_head_channels (:obj:`int`): The channels of policy head.
- fc_value_layers (:obj:`SequenceType`): The number of hidden layers used in value head (MLP head).
- fc_policy_layers (:obj:`SequenceType`): The number of hidden layers used in policy head (MLP head).
- value_head_hidden_channels (:obj:`SequenceType`): The number of hidden layers used in value head (MLP head).
- policy_head_hidden_channels (:obj:`SequenceType`): The number of hidden layers used in policy head (MLP head).
- output_support_size (:obj:`int`): The size of categorical value output.
- flatten_output_size_for_value_head (:obj:`int`): The size of flatten hidden states, i.e. the input size \
- flatten_input_size_for_value_head (:obj:`int`): The size of flatten hidden states, i.e. the input size \
of the value head.
- flatten_output_size_for_policy_head (:obj:`int`): The size of flatten hidden states, i.e. the input size \
- flatten_input_size_for_policy_head (:obj:`int`): The size of flatten hidden states, i.e. the input size \
of the policy head.
- last_linear_layer_init_zero (:obj:`bool`): Whether to use zero initializations for the last layer of \
value/policy mlp, default sets it to True.
Expand All @@ -255,8 +255,8 @@ def __init__(
"""
super().__init__()
self.continuous_action_space = continuous_action_space
self.flatten_output_size_for_value_head = flatten_output_size_for_value_head
self.flatten_output_size_for_policy_head = flatten_output_size_for_policy_head
self.flatten_input_size_for_value_head = flatten_input_size_for_value_head
self.flatten_input_size_for_policy_head = flatten_input_size_for_policy_head
self.norm_type = norm_type
self.sigma_type = sigma_type
self.fixed_sigma_value = fixed_sigma_value
Expand All @@ -274,13 +274,13 @@ def __init__(
self.conv1x1_policy = nn.Conv2d(num_channels, policy_head_channels, 1)
self.norm_value = nn.BatchNorm2d(value_head_channels)
self.norm_policy = nn.BatchNorm2d(policy_head_channels)
self.flatten_output_size_for_value_head = flatten_output_size_for_value_head
self.flatten_output_size_for_policy_head = flatten_output_size_for_policy_head
self.flatten_input_size_for_value_head = flatten_input_size_for_value_head
self.flatten_input_size_for_policy_head = flatten_input_size_for_policy_head
self.fc_value_head = MLP(
in_channels=self.flatten_output_size_for_value_head,
hidden_channels=fc_value_layers[0],
in_channels=self.flatten_input_size_for_value_head,
hidden_channels=value_head_hidden_channels[0],
out_channels=output_support_size,
layer_num=len(fc_value_layers) + 1,
layer_num=len(value_head_hidden_channels) + 1,
activation=activation,
norm_type='LN',
output_activation=False,
Expand All @@ -291,9 +291,9 @@ def __init__(
# sampled related core code
if self.continuous_action_space:
self.fc_policy_head = ReparameterizationHead(
input_size=self.flatten_output_size_for_policy_head,
input_size=self.flatten_input_size_for_policy_head,
output_size=action_space_size,
layer_num=len(fc_policy_layers) + 1,
layer_num=len(policy_head_hidden_channels) + 1,
sigma_type=self.sigma_type,
fixed_sigma_value=self.fixed_sigma_value,
activation=nn.ReLU(),
Expand All @@ -302,10 +302,10 @@ def __init__(
)
else:
self.fc_policy_head = MLP(
in_channels=self.flatten_output_size_for_policy_head,
hidden_channels=fc_policy_layers[0],
in_channels=self.flatten_input_size_for_policy_head,
hidden_channels=policy_head_hidden_channels[0],
out_channels=action_space_size,
layer_num=len(fc_policy_layers) + 1,
layer_num=len(policy_head_hidden_channels) + 1,
activation=activation,
norm_type='LN',
output_activation=False,
Expand Down Expand Up @@ -340,8 +340,8 @@ def forward(self, x: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
policy = self.norm_policy(policy)
policy = self.activation(policy)

value = value.reshape(-1, self.flatten_output_size_for_value_head)
policy = policy.reshape(-1, self.flatten_output_size_for_policy_head)
value = value.reshape(-1, self.flatten_input_size_for_value_head)
policy = policy.reshape(-1, self.flatten_input_size_for_policy_head)

value = self.fc_value_head(value)

Expand Down
Loading
Loading