Skip to content
New issue

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

Sign up for GitHub

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

Already on GitHub? Sign in to your account

[Cherry-pick for release] POCA Attention will use h_size for embedding size and not 128 (#5281) #5287

Merged
merged 1 commit into from
Apr 19, 2021
Merged

[Cherry-pick for release] POCA Attention will use h_size for embedding size and not 128 (#5281) #5287

Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
11 changes: 5 additions & 6 deletions ml-agents/mlagents/trainers/torch/networks.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -253,8 +253,6 @@ def forward(


class MultiAgentNetworkBody(torch.nn.Module):
ATTENTION_EMBEDDING_SIZE = 128

"""
A network body that uses a self attention layer to handle state
and action input from a potentially variable number of agents that
Expand Down Expand Up @@ -293,17 +291,18 @@ def __init__(
+ self.action_spec.continuous_size
)

attention_embeding_size = self.h_size
self.obs_encoder = EntityEmbedding(
obs_only_ent_size, None, self.ATTENTION_EMBEDDING_SIZE
obs_only_ent_size, None, attention_embeding_size
)
self.obs_action_encoder = EntityEmbedding(
q_ent_size, None, self.ATTENTION_EMBEDDING_SIZE
q_ent_size, None, attention_embeding_size
)

self.self_attn = ResidualSelfAttention(self.ATTENTION_EMBEDDING_SIZE)
self.self_attn = ResidualSelfAttention(attention_embeding_size)

self.linear_encoder = LinearEncoder(
self.ATTENTION_EMBEDDING_SIZE,
attention_embeding_size,
network_settings.num_layers,
self.h_size,
kernel_gain=(0.125 / self.h_size) ** 0.5,
Expand Down