PVNet 3.0.49__tar.gz → 3.0.51__tar.gz

{pvnet-3.0.49 → pvnet-3.0.51}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PVNet
-Version: 3.0.49
+Version: 3.0.51
 Summary: PVNet
 Author-email: Peter Dudfield <info@openclimatefix.org>
 Description-Content-Type: text/markdown

{pvnet-3.0.49 → pvnet-3.0.51}/PVNet.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: PVNet
-Version: 3.0.49
+Version: 3.0.51
 Summary: PVNet
 Author-email: Peter Dudfield <info@openclimatefix.org>
 Description-Content-Type: text/markdown

pvnet-3.0.51/pvnet/__init__.py

@@ -0,0 +1,2 @@
+"""PVNet"""
+__version__ = "3.0.51"

{pvnet-3.0.49 → pvnet-3.0.51}/pvnet/models/base_model.py

@@ -13,11 +13,10 @@ import torch
 import torch.nn.functional as F
 import wandb
 import yaml
-from huggingface_hub import ModelCard, ModelCardData, PyTorchModelHubMixin
+from huggingface_hub import ModelCard, ModelCardData
 from huggingface_hub.constants import CONFIG_NAME, PYTORCH_WEIGHTS_NAME
 from huggingface_hub.file_download import hf_hub_download
 from huggingface_hub.hf_api import HfApi
-from huggingface_hub.utils._deprecation import _deprecate_positional_args
 from ocf_datapipes.batch import BatchKey
 from ocf_ml_metrics.evaluation.evaluation import evaluation
 
@@ -141,33 +140,34 @@ def minimize_data_config(input_path, output_path, model):
         yaml.dump(config, outfile, default_flow_style=False)
 
 
-class PVNetModelHubMixin(PyTorchModelHubMixin):
+class PVNetModelHubMixin:
     """
     Implementation of [`PyTorchModelHubMixin`] to provide model Hub upload/download capabilities.
     """
 
     @classmethod
-    @_deprecate_positional_args(version="0.16")
-    def _from_pretrained(
+    def from_pretrained(
         cls,
         *,
         model_id: str,
         revision: str,
-        cache_dir: str,
-        force_download: bool,
-        proxies: Optional[Dict],
-        resume_download: bool,
-        local_files_only: bool,
-        token: Union[str, bool, None],
+        cache_dir: Optional[Union[str, Path]] = None,
+        force_download: bool = False,
+        proxies: Optional[Dict] = None,
+        resume_download: Optional[bool] = None,
+        local_files_only: bool = False,
+        token: Union[str, bool, None] = None,
         map_location: str = "cpu",
         strict: bool = False,
-        **model_kwargs,
     ):
         """Load Pytorch pretrained weights and return the loaded model."""
+
         if os.path.isdir(model_id):
             print("Loading weights from local directory")
             model_file = os.path.join(model_id, PYTORCH_WEIGHTS_NAME)
+            config_file = os.path.join(model_id, CONFIG_NAME)
         else:
+            # load model file
             model_file = hf_hub_download(
                 repo_id=model_id,
                 filename=PYTORCH_WEIGHTS_NAME,
@@ -180,11 +180,23 @@ class PVNetModelHubMixin(PyTorchModelHubMixin):
                 local_files_only=local_files_only,
             )
 
-        if "config" not in model_kwargs:
-            raise ValueError("Config must be supplied to instantiate model")
+            # load config file
+            config_file = hf_hub_download(
+                repo_id=model_id,
+                filename=CONFIG_NAME,
+                revision=revision,
+                cache_dir=cache_dir,
+                force_download=force_download,
+                proxies=proxies,
+                resume_download=resume_download,
+                token=token,
+                local_files_only=local_files_only,
+            )
+
+        with open(config_file, "r", encoding="utf-8") as f:
+            config = json.load(f)
 
-        model_kwargs.update(model_kwargs.pop("config"))
-        model = hydra.utils.instantiate(model_kwargs)
+        model = hydra.utils.instantiate(config)
 
         state_dict = torch.load(model_file, map_location=torch.device(map_location))
         model.load_state_dict(state_dict, strict=strict)  # type: ignore

{pvnet-3.0.49 → pvnet-3.0.51}/pvnet/models/multimodal/multimodal.py

@@ -48,6 +48,7 @@ class Model(MultimodalBaseModel):
         add_image_embedding_channel: bool = False,
         include_gsp_yield_history: bool = True,
         include_sun: bool = True,
+        include_time: bool = False,
         embedding_dim: Optional[int] = 16,
         forecast_minutes: int = 30,
         history_minutes: int = 60,
@@ -97,6 +98,7 @@ class Model(MultimodalBaseModel):
                 embedding of the GSP ID.
             include_gsp_yield_history: Include GSP yield data.
             include_sun: Include sun azimuth and altitude data.
+            include_time: Include sine and cosine of dates and times.
             embedding_dim: Number of embedding dimensions to use for GSP ID. Not included if set to
                 `None`.
             forecast_minutes: The amount of minutes that should be forecasted.
@@ -141,6 +143,7 @@ class Model(MultimodalBaseModel):
         self.include_nwp = nwp_encoders_dict is not None and len(nwp_encoders_dict) != 0
         self.include_pv = pv_encoder is not None
         self.include_sun = include_sun
+        self.include_time = include_time
         self.include_wind = wind_encoder is not None
         self.include_sensor = sensor_encoder is not None
         self.embedding_dim = embedding_dim
@@ -283,6 +286,16 @@ class Model(MultimodalBaseModel):
             # Update num features
             fusion_input_features += 16
 
+        if self.include_time:
+            self.time_fc1 = nn.Linear(
+                in_features=4
+                * (self.forecast_len + self.forecast_len_ignore + self.history_len + 1),
+                out_features=32,
+            )
+
+            # Update num features
+            fusion_input_features += 32
+
         if include_gsp_yield_history:
             # Update num features
             fusion_input_features += self.history_len

{pvnet-3.0.49 → pvnet-3.0.51}/pvnet/models/utils.py

@@ -149,7 +149,8 @@ class WeightedLosses:
             self.decay_rate = math.log(2)
 
         # make weights from decay rate
-        weights = torch.from_numpy(np.exp(-self.decay_rate * np.arange(self.forecast_length)))
+        weights = np.exp(-self.decay_rate * np.arange(self.forecast_length))
+        weights = torch.tensor(weights)
 
         # normalized the weights, so there mean is 1.
         # To calculate the loss, we times the weights by the differences between truth

pvnet-3.0.49/pvnet/__init__.py

@@ -1,2 +0,0 @@
-"""PVNet"""
-__version__ = "3.0.49"