Fixing the bug in raindrop when running on multiple devices

.github/workflows/testing_ci.yml

@@ -23,7 +23,8 @@ jobs:
                 python-version: ["3.7", "3.8", "3.9", "3.10"]
 
         steps:
-            - uses: actions/checkout@v3
+            - name: Check out the repo code
+              uses: actions/checkout@v3
 
             - name: Set up Conda
               uses: conda-incubator/setup-miniconda@v2

README.md

@@ -61,8 +61,8 @@ with the missing parts in their data. PyPOTS will keep integrating classical and
 algorithms for partially-observed multivariate time series. For sure, besides various algorithms, PyPOTS is going to
 have unified APIs together with detailed documentation and interactive examples across algorithms as tutorials.
 
-👍 **Please** star this repo to help others notice PyPOTS if you think it is a useful toolkit.
-**Please** properly [cite PyPOTS](https://github.com/WenjieDu/PyPOTS#-citing-pypots) in your publications 
+🤗 **Please** star this repo to help others notice PyPOTS if you think it is a useful toolkit.
+**Please** properly [cite PyPOTS](https://github.com/WenjieDu/PyPOTS#-citing-pypots) in your publications
 if it helps with your research. This really means a lot to our open-source research. Thank you!
 
 <a href="https://github.com/WenjieDu/TSDB">
@@ -177,8 +177,8 @@ PyPOTS supports imputation, classification, clustering, and forecasting tasks on
 
 
 ## ❖ Citing PyPOTS
-**[Updates in Jun 2023]** 🎉A short version of the PyPOTS paper is accepted by the 9th SIGKDD international workshop on 
-Mining and Learning from Time Series ([MiLeTS'23](https://kdd-milets.github.io/milets2023/))). 
+**[Updates in Jun 2023]** 🎉A short version of the PyPOTS paper is accepted by the 9th SIGKDD international workshop on
+Mining and Learning from Time Series ([MiLeTS'23](https://kdd-milets.github.io/milets2023/))).
 Besides, PyPOTS has been included as a [PyTorch Ecosystem](https://pytorch.org/ecosystem/) project.
 
 The paper introducing PyPOTS is available on arXiv at [this URL](https://arxiv.org/abs/2305.18811),
@@ -266,6 +266,6 @@ PyPOTS community is open, transparent, and surely friendly. Let's work together
 <details>
 <summary>🏠 Visits</summary>
 <a href="https://github.com/WenjieDu/PyPOTS">
-    <img alt="PyPOTS visits" align="left" src="https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fgithub.jparrowsec.cn%2FPyPOTS%2FPyPOTS&count_bg=%23009A0A&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=Hits&edge_flat=false">
+    <img alt="PyPOTS visits" align="left" src="https://hits.seeyoufarm.com/api/count/incr/badge.svg?url=https%3A%2F%2Fgithub.jparrowsec.cn%2FPyPOTS%2FPyPOTS&count_bg=%23009A0A&title_bg=%23555555&icon=&icon_color=%23E7E7E7&title=Visits%20since%20May%202022&edge_flat=false">
 </a>
 </details>

docs/about_us.rst

@@ -19,8 +19,12 @@ Maciej Skrabski
 
 All Contributors
 """"""""""""""""
-
-PyPOTS exists thanks to `all the nice people <https://github.com/WenjieDu/PyPOTS/graphs/contributors>`_ who contribute their time to work on the project:
+PyPOTS exists thanks to all the nice people who contribute their time to work on the project (including the repositories
+`PyPOTS <https://github.com/WenjieDu/PyPOTS/graphs/contributors>`_,
+`BrewPOTS <https://github.com/WenjieDu/BrewPOTS/graphs/contributors>`_,
+`TSDB <https://github.com/WenjieDu/TSDB/graphs/contributors>`_,
+`PyCorruptor <https://github.com/WenjieDu/PyCorruptor/graphs/contributors>`_
+):
 
 .. raw:: html
 

docs/faq.rst

@@ -14,7 +14,7 @@ Note this exception only applies if you commit to the maintenance of your model
 
 Becoming a Maintainer
 ^^^^^^^^^^^^^^^^^^^^^
-To become a maintainer of PyPOTS, you should
+To join the team and become a maintainer of PyPOTS, you should
 
 1. be active on GitHub and watch PyPOTS repository to receive the latest news from it;
 2. be familiar with the PyPOTS codebase and have made at least one pull request merged into branch ``main`` of PyPOTS,
@@ -31,6 +31,10 @@ Once you obtain the role, you'll be listed as a member on the ``About Us`` pages
 and
 `PyPOTS docs site <https://docs.pypots.com/en/latest/about_us.html>`_.
 
+**NOTE**: The maintainer role is not permanent. The role is called "maintainer" because it actively maintains the project.
+You can take a leave of absence from the role with notice at any time.
+But if you're inactive for a long time (more than three months. With reasons, a longer period is allowed for sure), your role will be deactivated.
+
 
 Our Development Principles
 ^^^^^^^^^^^^^^^^^^^^^^^^^^

docs/index.rst

@@ -77,7 +77,7 @@ Welcome to PyPOTS docs!
 ⦿ `Mission`: PyPOTS is born to become a handy toolbox that is going to make data mining on POTS easy rather than tedious, to help engineers and researchers focus more on the core problems in their hands rather than on how to deal with the missing parts in their data. PyPOTS will keep integrating classical and the latest state-of-the-art data mining algorithms for partially-observed multivariate time series. For sure, besides various algorithms, PyPOTS is going to have unified APIs together with detailed documentation and interactive examples across algorithms as tutorials.
 
 👍 **Please** star this repo to help others notice PyPOTS if you think it is a useful toolkit.
-**Please** properly `cite PyPOTS <https://docs.pypots.com/en/latest/index.html#id14>`_ in your publications
+**Please** properly `cite PyPOTS <https://docs.pypots.com/en/latest/milestones.html#citing-pypots>`_ in your publications
 if it helps with your research. This really means a lot to our open-source research. Thank you!
 
 .. image:: https://raw.githubusercontent.com/WenjieDu/TSDB/main/docs/_static/figs/TSDB_logo.svg?sanitize=true

docs/milestones.rst

@@ -3,8 +3,9 @@ Citation and Milestones
 
 Citing PyPOTS
 ^^^^^^^^^^^^^
-**[Updates in Jun 2023]** 🎉A short version of the PyPOTS paper is accepted by the 9th SIGKDD International Workshop on
+**[Updates in Jun 2023]** 🎉A short version of the PyPOTS paper is accepted by the 9th SIGKDD international workshop on
 Mining and Learning from Time Series (`MiLeTS'23 <https://kdd-milets.github.io/milets2023/>`_).
+Besides, PyPOTS has been included as a `PyTorch Ecosystem <https://pytorch.org/ecosystem/>`_ project.
 
 PyPOTS paper is available on arXiv at `this URL <https://arxiv.org/abs/2305.18811>`_.,
 and we are pursuing to publish it in prestigious academic venues, e.g. JMLR (track for
@@ -48,4 +49,4 @@ Project Milestones
 - 2023-05: PyPOTS v0.1 is released, and `the preprint paper <https://arxiv.org/abs/2305.18811>`_ is published on arXiv;
 - 2023-06: A short version of PyPOTS paper is accepted by the 9th SIGKDD International
   Workshop on Mining and Learning from Time Series (`MiLeTS'23 <https://kdd-milets.github.io/milets2023/>`_);
-- 2023-07: PyPOTS has been accepted as a `PyTorch Ecosystem <https://pytorch.org/ecosystem/>`_ project;
+- 2023-07: PyPOTS has been accepted as a `PyTorch Ecosystem <https://pytorch.org/ecosystem/>`_ project;

pypots/classification/raindrop/model.py

@@ -77,7 +77,6 @@ def __init__(
         self.device = device
 
         # create modules
-        self.global_structure = torch.ones(n_features, n_features, device=self.device)
         if self.static:
             self.static_emb = nn.Linear(d_static, n_features)
         else:
@@ -101,8 +100,6 @@ def __init__(
             )
         self.transformer_encoder = TransformerEncoder(encoder_layers, n_layers)
 
-        self.adj = torch.ones([self.n_features, self.n_features], device=self.device)
-
         self.R_u = Parameter(torch.Tensor(1, self.n_features * self.d_ob))
 
         self.ob_propagation = ObservationPropagation(
@@ -152,28 +149,28 @@ def classify(self, inputs: dict) -> torch.Tensor:
         -------
         prediction : torch.Tensor
         """
-        src = inputs["X"]
+        src = inputs["X"].permute(1, 0, 2)
         static = inputs["static"]
-        times = inputs["timestamps"]
+        times = inputs["timestamps"].permute(1, 0)
         lengths = inputs["lengths"]
-        missing_mask = inputs["missing_mask"]
+        missing_mask = inputs["missing_mask"].permute(1, 0, 2)
 
         max_len, batch_size = src.shape[0], src.shape[1]
 
         src = torch.repeat_interleave(src, self.d_ob, dim=-1)
         h = F.relu(src * self.R_u)
-        pe = self.pos_encoder(times).to(self.device)
+        pe = self.pos_encoder(times).to(src.device)
         if static is not None:
             emb = self.static_emb(static)
 
         h = self.dropout(h)
 
         mask = torch.arange(max_len)[None, :] >= (lengths.cpu()[:, None])
-        mask = mask.squeeze(1).to(self.device)
+        mask = mask.squeeze(1).to(src.device)
 
         x = h
 
-        adj = self.global_structure
+        adj = torch.ones(self.n_features, self.n_features, device=src.device)
         adj[torch.eye(self.n_features, dtype=torch.bool)] = 1
 
         edge_index = torch.nonzero(adj).T
@@ -182,10 +179,10 @@ def classify(self, inputs: dict) -> torch.Tensor:
         batch_size = src.shape[1]
         n_step = src.shape[0]
         output = torch.zeros(
-            [n_step, batch_size, self.n_features * self.d_ob], device=self.device
+            [n_step, batch_size, self.n_features * self.d_ob], device=src.device
         )
 
-        alpha_all = torch.zeros([edge_index.shape[1], batch_size], device=self.device)
+        alpha_all = torch.zeros([edge_index.shape[1], batch_size], device=src.device)
 
         # iterate on each sample
         for unit in range(0, batch_size):
@@ -240,13 +237,11 @@ def classify(self, inputs: dict) -> torch.Tensor:
 
         r_out = self.transformer_encoder(output, src_key_padding_mask=mask)
 
-        sensor_wise_mask = self.sensor_wise_mask
-
-        lengths2 = lengths.unsqueeze(1).to(self.device)
+        lengths2 = lengths.unsqueeze(1).to(src.device)
         mask2 = mask.permute(1, 0).unsqueeze(2).long()
-        if sensor_wise_mask:
+        if self.sensor_wise_mask:
             output = torch.zeros(
-                [batch_size, self.n_features, self.d_ob + 16], device=self.device
+                [batch_size, self.n_features, self.d_ob + 16], device=src.device
             )
             extended_missing_mask = missing_mask.view(-1, batch_size, self.n_features)
             for se in range(self.n_features):
@@ -458,10 +453,6 @@ def _assemble_input_for_training(self, data: dict) -> dict:
         lengths = torch.tensor([n_steps] * bz, dtype=torch.float)
         times = torch.tensor(range(n_steps), dtype=torch.float).repeat(bz, 1)
 
-        X = X.permute(1, 0, 2)
-        missing_mask = missing_mask.permute(1, 0, 2)
-        times = times.permute(1, 0)
-
         inputs = {
             "X": X,
             "static": None,
@@ -488,10 +479,6 @@ def _assemble_input_for_testing(self, data: dict) -> dict:
         lengths = torch.tensor([n_steps] * bz, dtype=torch.float)
         times = torch.tensor(range(n_steps), dtype=torch.float).repeat(bz, 1)
 
-        X = X.permute(1, 0, 2)
-        missing_mask = missing_mask.permute(1, 0, 2)
-        times = times.permute(1, 0)
-
         inputs = {
             "X": X,
             "static": None,

pypots/classification/raindrop/modules.py

@@ -66,7 +66,9 @@ def forward(self, time_vectors: torch.Tensor) -> torch.Tensor:
         timescales = self.max_len ** np.linspace(0, 1, self._num_timescales)
 
         times = time_vectors.unsqueeze(2)
-        scaled_time = times / torch.Tensor(timescales[None, None, :])
+        scaled_time = times / torch.from_numpy(timescales[None, None, :]).to(
+            time_vectors.device
+        )
         pe = torch.cat(
             [torch.sin(scaled_time), torch.cos(scaled_time)], dim=-1
         )  # T x B x d_model