.vscode/launch.json

@@ -2097,6 +2097,22 @@
             "program": "run.py",
             "console": "integratedTerminal",
             "args": "--config ./config/iTransformer/METR-LA.yaml"
-        }
+        },
+        {
+            "name": "iTransformer: AirQuality",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "run.py",
+            "console": "integratedTerminal",
+            "args": "--config ./config/iTransformer/AirQuality.yaml"
+        },
+        {
+            "name": "iTransformer: SolarEnergy",
+            "type": "debugpy",
+            "request": "launch",
+            "program": "run.py",
+            "console": "integratedTerminal",
+            "args": "--config ./config/iTransformer/SolarEnergy.yaml"
+        },
     ]
 }

config/iTransformer/AirQuality.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: AirQuality
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 16
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 6
+  lag: 24
+  normalizer: std
+  num_nodes: 35
+  steps_per_day: 24
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 35
+  plot: false
+  real_value: true
+  weight_decay: 0

config/iTransformer/BJTaxi-Inflow.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: BJTaxi-InFlow
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 32
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 1
+  lag: 24
+  normalizer: std
+  num_nodes: 1024
+  steps_per_day: 48
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 1024
+  plot: false
+  real_value: true
+  weight_decay: 0

config/iTransformer/BJTaxi-Outflow.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: BJTaxi-OutFlow
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 32
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 1
+  lag: 24
+  normalizer: std
+  num_nodes: 1024
+  steps_per_day: 48
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 1024
+  plot: false
+  real_value: true
+  weight_decay: 0

config/iTransformer/METR-LA.yaml

@@ -1,6 +1,6 @@
 basic:
   dataset: METR-LA
-  device: cuda:0
+  device: cuda:1
   mode: train
   model: iTransformer
   seed: 2023

config/iTransformer/NYCBike-Inflow.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: NYCBike-InFlow
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 32
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 1
+  lag: 24
+  normalizer: std
+  num_nodes: 128
+  steps_per_day: 48
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 128
+  plot: false
+  real_value: true
+  weight_decay: 0

config/iTransformer/NYCBike-Outflow.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: NYCBike-OutFlow
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 32
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 1
+  lag: 24
+  normalizer: std
+  num_nodes: 128
+  steps_per_day: 48
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 128
+  plot: false
+  real_value: true
+  weight_decay: 0

config/iTransformer/PEMS-BAY.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: PEMS-BAY
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 16
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 1
+  lag: 24
+  normalizer: std
+  num_nodes: 325
+  steps_per_day: 288
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 325
+  plot: false
+  real_value: true
+  weight_decay: 0

config/iTransformer/SolarEnergy.yaml

@@ -0,0 +1,52 @@
+basic:
+  dataset: SolarEnergy
+  device: cuda:0
+  mode: train
+  model: iTransformer
+  seed: 2023
+
+data:
+  batch_size: 16
+  column_wise: false
+  days_per_week: 7
+  horizon: 24
+  input_dim: 6
+  lag: 24
+  normalizer: std
+  num_nodes: 137
+  steps_per_day: 24
+  test_ratio: 0.2
+  val_ratio: 0.2
+
+model:
+  activation: gelu
+  seq_len: 24
+  pred_len: 24
+  d_model: 128
+  d_ff: 2048
+  dropout: 0.1
+  e_layers: 2
+  n_heads: 8
+  output_attention: False
+
+
+train:
+  batch_size: 16
+  debug: false
+  early_stop: true
+  early_stop_patience: 15
+  epochs: 100
+  grad_norm: false
+  log_step: 1000
+  loss_func: mae
+  lr_decay: true
+  lr_decay_rate: 0.3
+  lr_decay_step: 5,20,40,70
+  lr_init: 0.0001
+  mae_thresh: None
+  mape_thresh: 0.001
+  max_grad_norm: 5
+  output_dim: 137
+  plot: false
+  real_value: true
+  weight_decay: 0

dataloader/TSloader.py

@@ -7,6 +7,7 @@ import torch
 
 def get_dataloader(args, normalizer="std", single=True):
     data = load_st_dataset(args)
+    data = data[..., 0:1]
 
     args = args["data"]
     L, N, F = data.shape