Kumo.ai home page
Search or ask...
Community
Contact Support
Start for free
Start for free
Search...
Navigation
Graph Transformers
num_heads
Guide
Examples
Reference
Troubleshooting
Releases
Predictive Query
Overview
Primary Commands
Aggregation Operators
Boolean Operators
Model Plan
Column Processing
Model Architecture
GNN
Graph Transformers
channels
num_layer
num_heads
dropout
positional_encodings
activation
distance_measure
handle_new_target_entities
module
normalization
num_post_message_passing_layers
num_pre_message_passing_layers
ranking_embedding_loss_coeff
output_embedding_dim
prediction_time_encodings
target_embedding_mode
use_seq_id
Neighbor Sampling
Optimization
Training Job Plan
Training Table Generation
SDK
SDK Reference
Graph Transformers
num_heads
num_heads: <list[int]>
(Optional)
Description
Specifies the number of attention heads in the GraphTransformer model.
Supported Task Types
Node classification and regression
Default Values
run_mode
Default Value
FAST
[8]
NORMAL
[8]
BEST
[8]
num_layer
dropout
On this page
num_heads: <list[int]> (Optional)
Description
Supported Task Types
Default Values
