Merge pull request #427 from WenjieDu/dev

Fix ETSformer tuning bug, and release v0.6rc1

pypots/__init__.py

@@ -22,7 +22,7 @@
 #
 # Dev branch marker is: 'X.Y.dev' or 'X.Y.devN' where N is an integer.
 # 'X.Y.dev0' is the canonical version of 'X.Y.dev'
-__version__ = "0.5"
+__version__ = "0.6rc1"
 
 
 from . import imputation, classification, clustering, forecasting, optim, data, utils

pypots/cli/tuning.py

@@ -14,6 +14,7 @@
 from .base import BaseCommand
 from .utils import load_package_from_path
 from ..classification import BRITS as BRITS_classification
+from ..classification import GRUD as GRUD_classification
 from ..classification import Raindrop
 from ..clustering import CRLI, VaDER
 from ..data.saving.h5 import load_dict_from_h5
@@ -80,8 +81,9 @@
  "pypots.imputation.GPVAE": GPVAE,
  "pypots.imputation.BRITS": BRITS,
  "pypots.imputation.MRNN": MRNN,
+ "pypots.imputation.GRUD": GRUD,
  # classification models
- "pypots.classification.GRUD": GRUD,
+ "pypots.classification.GRUD": GRUD_classification,
  "pypots.classification.BRITS": BRITS_classification,
  "pypots.classification.Raindrop": Raindrop,
  # clustering models
@@ -248,7 +250,7 @@ def run(self):
  )
  raise RuntimeError(
  f"Hyperparameters do not match. Mismatched hyperparameters "
- f"(in the tuning configuration but not in the given model's arguments): {list(mismatched)}"
+ f"(in the tuning configuration but not in {model_class.__name__}'s arguments): {list(mismatched)}"
  )
 
  # initializing optimizer and model

pypots/imputation/etsformer/model.py

@@ -103,14 +103,14 @@ class ETSformer(BaseNNImputer):
 
  def __init__(
  self,
- n_steps,
- n_features,
- n_e_layers,
- n_d_layers,
- d_model,
- n_heads,
- d_ffn,
- top_k,
+ n_steps: int,
+ n_features: int,
+ n_e_layers: int,
+ n_d_layers: int,
+ d_model: int,
+ n_heads: int,
+ d_ffn: int,
+ top_k: int,
  dropout: float = 0,
  ORT_weight: float = 1,
  MIT_weight: float = 1,

pypots/imputation/fedformer/model.py

@@ -111,13 +111,13 @@ class FEDformer(BaseNNImputer):
 
  def __init__(
  self,
- n_steps,
- n_features,
- n_layers,
- d_model,
- n_heads,
- d_ffn,
- moving_avg_window_size,
+ n_steps: int,
+ n_features: int,
+ n_layers: int,
+ d_model: int,
+ n_heads: int,
+ d_ffn: int,
+ moving_avg_window_size: int,
  dropout: float = 0,
  version="Fourier",
  modes=32,

pypots/nn/modules/etsformer/layers.py

@@ -160,7 +160,7 @@ def forward(self, x):
  f = fft.rfftfreq(t)[self.low_freq :]
 
  x_freq, index_tuple = self.topk_freq(x_freq)
- f = repeat(f, "f -> b f d", b=x_freq.size(0), d=x_freq.size(2))
+ f = repeat(f, "f -> b f d", b=x_freq.size(0), d=x_freq.size(2)).to(x_freq.device)
  f = rearrange(f[index_tuple], "b f d -> b f () d").to(x_freq.device)
 
  return self.extrapolate(x_freq, f, t)