gcpnet_cpd.yaml

# @package _global_

# to execute this experiment run:
# python train.py experiment=gcpnet_cpd

defaults:
  - override /datamodule: cath_cpd.yaml
  - override /model: gcpnet_cpd.yaml
  - override /callbacks: default.yaml
  - override /trainer: default.yaml

# all parameters below will be merged with parameters from default configurations set above
# this allows you to overwrite only specified parameters

tags: ["gcpnet", "cpd"]

seed: 42

callbacks:
  model_checkpoint:
    monitor: "val/loss"
    save_top_k: 30
  early_stopping:
    monitor: "val/loss"

trainer:
  min_epochs: 100
  max_epochs: 1000
  strategy: ddp_find_unused_parameters_false
  accelerator: gpu
  devices: 1
  num_nodes: 1
  accumulate_grad_batches: 4

model:
  optimizer:
    lr: 1e-4
    weight_decay: 1e-8
  
  model_cfg:
    h_hidden_dim: 100
    chi_hidden_dim: 16
    e_hidden_dim: 32
    xi_hidden_dim: 4

    num_encoder_layers: 9
    num_decoder_layers: 3
    dropout: 0.2

    decoder_residual_updates: true

  module_cfg:
    selected_GCP:
      # which version of the GCP module to use (e.g., GCP or GCP2)
      _target_: src.models.components.gcpnet.GCP2
      _partial_: true

    norm_x_diff: true

    scalar_gate: 0
    vector_gate: true  # note: For both GCP and GCP2, this parameter is used;   For GCP2, this mimics updating vector features without directly using frame vectors
    vector_residual: false  # note: For both GCP and GCP2, this parameter is used
    vector_frame_residual: false  # note: for GCP2, this parameter is unused
    frame_gate: false  # note: for GCP2, if this parameter and `vector_gate` are both set to `false`, row-wise vector self-gating is applied instead
    sigma_frame_gate: false  # note: For GCP, this parameter overrides `frame_gate`;   For GCP2, this parameter is unused and is replaced in functionality by `vector_gate`

    scalar_nonlinearity: relu
    vector_nonlinearity: 

    nonlinearities:
      - ${..scalar_nonlinearity}
      - ${..vector_nonlinearity}

    bottleneck: 4

    vector_linear: true
    vector_identity: true

    default_vector_residual: false
    default_bottleneck: 4

    ablate_frame_updates: false
    ablate_scalars: false
    ablate_vectors: false

  layer_cfg:
    mp_cfg:
      edge_encoder: false
      edge_gate: false
      num_message_layers: 8
      message_residual: 0
      message_ff_multiplier: 1
      self_message: true
      use_residual_message_gcp: true

    pre_norm: false
    num_feedforward_layers: 2
    dropout: 0.1

    nonlinearity_slope: 1e-2

datamodule:
  batch_size: 8
  num_workers: 4

logger:
  wandb:
    name: 12282022_GCPNet
    group: "CPD"
    tags: ${tags}