add and test extra_run_config: dict | None = None keyword to Trainer …

…to specify run params like batch_size that aren't already recorded by the trainer_args dict

chgnet/trainer/trainer.py

@@ -57,7 +57,8 @@ def __init__(
         use_device: str | None = None,
         check_cuda_mem: bool = False,
         wandb_path: str | None = None,
-        wandb_kwargs: dict | None = None,
+        wandb_init_kwargs: dict | None = None,
+        extra_run_config: dict | None = None,
         **kwargs,
     ) -> None:
         """Initialize all hyper-parameters for trainer.
@@ -99,7 +100,11 @@ def __init__(
             wandb_path (str | None): The project and run name separated by a slash:
                 "project/run_name". If None, wandb logging is not used.
                 Default = None
-            wandb_kwargs (dict): additional kwargs for wandb.init. Default = None
+            wandb_init_kwargs (dict): Additional kwargs to pass to wandb.init.
+                Default = None
+            extra_run_config (dict): Additional hyper-params to be recorded by wandb
+                that are not included in the trainer_args. Default = None
+
             **kwargs (dict): additional hyper-params for optimizer, scheduler, etc.
         """
         # Store trainer args for reproducibility
@@ -213,12 +218,18 @@ def __init__(
                     "Weights and Biases not installed. pip install wandb to use "
                     "wandb logging."
                 )
-            project, run_name = wandb_path.split("/")
+            if wandb_path.count("/") == 1:
+                project, run_name = wandb_path.split("/")
+            else:
+                raise ValueError(
+                    f"{wandb_path=} should be in the format 'project/run_name' "
+                    "(no extra slashes)"
+                )
             wandb.init(
                 project=project,
                 name=run_name,
-                config=self.trainer_args,
-                **(wandb_kwargs or {}),
+                config=self.trainer_args | (extra_run_config or {}),
+                **(wandb_init_kwargs or {}),
             )
 
     def train(

tests/test_trainer.py

@@ -7,6 +7,7 @@
 import torch
 from pymatgen.core import Lattice, Structure
 
+import wandb
 from chgnet.data.dataset import StructureData, get_train_val_test_loader
 from chgnet.model import CHGNet
 from chgnet.trainer import Trainer
@@ -42,16 +43,19 @@ def test_trainer(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> None:
     train_loader, val_loader, _test_loader = get_train_val_test_loader(
         data, batch_size=16, train_ratio=0.9, val_ratio=0.05
     )
+    extra_run_config = dict(some_other_hyperparam=42)
     trainer = Trainer(
         model=chgnet,
         targets="efsm",
         optimizer="Adam",
         criterion="MSE",
         learning_rate=1e-2,
         epochs=5,
-        wandb_path="",
-        wandb_kwargs=dict(anonymous="allow"),
+        wandb_path="test/run",
+        wandb_init_kwargs=dict(anonymous="must"),
+        extra_run_config=extra_run_config,
     )
+    assert dict(wandb.config).items() >= extra_run_config.items()
     dir_name = "test_tmp_dir"
     test_dir = tmp_path / dir_name
     trainer.train(train_loader, val_loader, save_dir=test_dir)
@@ -70,7 +74,7 @@ def test_trainer(tmp_path: Path, monkeypatch: pytest.MonkeyPatch) -> None:
     err_msg = "Weights and Biases not installed. pip install wandb to use wandb logging"
     with monkeypatch.context() as ctx, pytest.raises(ImportError, match=err_msg):  # noqa: PT012
         ctx.setattr("chgnet.trainer.trainer.wandb", None)
-        _ = Trainer(model=chgnet, wandb_path="radicalai/chgnet-test-finetune")
+        _ = Trainer(model=chgnet, wandb_path="some-org/some-project")
 
 
 def test_trainer_composition_model(tmp_path: Path) -> None: