Merge branch 'main' into composition-model

MANIFEST.in

@@ -1,7 +1,7 @@
 graft src
 
 include LICENSE
-include README.md
+include README.rst
 
 prune developer
 prune docs

pyproject.toml

@@ -3,7 +3,7 @@ name = "metatrain"
 dynamic = ["version"]
 requires-python = ">=3.9"
 
-readme = "README.md"
+readme = "README.rst"
 license = {text = "BSD-3-Clause"}
 description = "Training and evaluating machine learning models for atomistic systems."
 authors = [{name = "metatrain developers"}]
@@ -12,7 +12,7 @@ dependencies = [
     "ase < 3.23.0",
     "metatensor-learn==0.2.3",
     "metatensor-operations==0.2.3",
-    "metatensor-torch==0.5.4",
+    "metatensor-torch==0.5.5",
     "jsonschema",
     "omegaconf",
     "python-hostlist",

src/metatrain/experimental/alchemical_model/model.py

@@ -129,7 +129,7 @@ def forward(
     def load_checkpoint(cls, path: Union[str, Path]) -> "AlchemicalModel":
 
         # Load the checkpoint
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         model_hypers = checkpoint["model_hypers"]
         model_state_dict = checkpoint["model_state_dict"]
 

src/metatrain/experimental/alchemical_model/trainer.py

@@ -349,7 +349,7 @@ def save_checkpoint(self, model, path: Union[str, Path]):
     def load_checkpoint(cls, path: Union[str, Path], train_hypers) -> "Trainer":
 
         # Load the checkpoint
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         model_hypers = checkpoint["model_hypers"]
         model_state_dict = checkpoint["model_state_dict"]
         epoch = checkpoint["epoch"]

src/metatrain/experimental/pet/model.py

@@ -114,14 +114,14 @@ def forward(
     @classmethod
     def load_checkpoint(cls, path: Union[str, Path]) -> "PET":
 
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         hypers = checkpoint["hypers"]
         dataset_info = checkpoint["dataset_info"]
         model = cls(
             model_hypers=hypers["ARCHITECTURAL_HYPERS"], dataset_info=dataset_info
         )
 
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         state_dict = checkpoint["checkpoint"]["model_state_dict"]
 
         ARCHITECTURAL_HYPERS = Hypers(model.hypers)

src/metatrain/experimental/pet/trainer.py

@@ -163,7 +163,7 @@ def train(
         else:
             load_path = self.pet_dir / "best_val_rmse_energies_model_state_dict"
 
-        state_dict = torch.load(load_path)
+        state_dict = torch.load(load_path, weights_only=False)
 
         ARCHITECTURAL_HYPERS = Hypers(model.hypers)
         raw_pet = PET(ARCHITECTURAL_HYPERS, 0.0, len(model.atomic_types))
@@ -186,7 +186,7 @@ def save_checkpoint(self, model, path: Union[str, Path]):
         # together with the hypers inside a file that will act as a metatrain
         # checkpoint
         checkpoint_path = self.pet_dir / "checkpoint"  # type: ignore
-        checkpoint = torch.load(checkpoint_path)
+        checkpoint = torch.load(checkpoint_path, weights_only=False)
         torch.save(
             {
                 "checkpoint": checkpoint,
@@ -204,7 +204,7 @@ def load_checkpoint(cls, path: Union[str, Path], train_hypers) -> "Trainer":
         # This function loads a metatrain PET checkpoint and returns a Trainer
         # instance with the hypers, while also saving the checkpoint in the
         # class
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         trainer = cls(train_hypers)
         trainer.pet_checkpoint = checkpoint["checkpoint"]
         return trainer

src/metatrain/experimental/soap_bpnn/model.py

@@ -290,7 +290,7 @@ def forward(
     def load_checkpoint(cls, path: Union[str, Path]) -> "SoapBpnn":
 
         # Load the checkpoint
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         model_hypers = checkpoint["model_hypers"]
         model_state_dict = checkpoint["model_state_dict"]
 

src/metatrain/experimental/soap_bpnn/trainer.py

@@ -25,13 +25,6 @@
 logger = logging.getLogger(__name__)
 
 
-# Filter out the second derivative and device warnings from rascaline-torch
-warnings.filterwarnings("ignore", category=UserWarning, message="second derivative")
-warnings.filterwarnings(
-    "ignore", category=UserWarning, message="Systems data is on device"
-)
-
-
 class Trainer:
     def __init__(self, train_hypers):
         self.hypers = train_hypers
@@ -48,6 +41,17 @@ def train(
         val_datasets: List[Union[Dataset, torch.utils.data.Subset]],
         checkpoint_dir: str,
     ):
+        # Filter out the second derivative and device warnings from rascaline
+        warnings.filterwarnings(action="ignore", message="Systems data is on device")
+        warnings.filterwarnings(
+            action="ignore",
+            message="second derivatives with respect to positions are not implemented",
+        )
+        warnings.filterwarnings(
+            action="ignore",
+            message="second derivatives with respect to cell matrix",
+        )
+
         assert dtype in SoapBpnn.__supported_dtypes__
 
         is_distributed = self.hypers["distributed"]
@@ -248,6 +252,7 @@ def train(
                 targets = average_by_num_atoms(targets, systems, per_structure_targets)
 
                 train_loss_batch = loss_fn(predictions, targets)
+
                 train_loss_batch.backward()
                 optimizer.step()
 
@@ -368,7 +373,7 @@ def save_checkpoint(self, model, path: Union[str, Path]):
     def load_checkpoint(cls, path: Union[str, Path], train_hypers) -> "Trainer":
 
         # Load the checkpoint
-        checkpoint = torch.load(path)
+        checkpoint = torch.load(path, weights_only=False)
         model_hypers = checkpoint["model_hypers"]
         model_state_dict = checkpoint["model_state_dict"]
         epoch = checkpoint["epoch"]

tests/cli/test_train_model.py

@@ -449,8 +449,8 @@ def test_model_consistency_with_seed(options, monkeypatch, tmp_path, seed):
 
     train_model(options, output="model2.pt")
 
-    m1 = torch.load("model1.ckpt")
-    m2 = torch.load("model2.ckpt")
+    m1 = torch.load("model1.ckpt", weights_only=False)
+    m2 = torch.load("model2.ckpt", weights_only=False)
 
     for i in m1["model_state_dict"]:
         tensor1 = m1["model_state_dict"][i]