Merge pull request #147 from cnellington/dev

Bugfixes for easy modules predict_params, added early stopping kwargs to easy modules.

README.md

@@ -81,6 +81,7 @@ Many people have helped. Check out [ACKNOWLEDGEMENTS.md](https://github.com/cnel
 
 
 ## Videos
+- [Cold Spring Harbor Laboratory: Contextualized Graphical Models Reveal Sample-Specific Transcriptional Networks for 7000 Tumors](https://www.youtube.com/watch?v=MTcjFK-YwCw)
 - [Sample-Specific Models for Interpretable Analysis with Applications to Disease Subtyping](http://www.birs.ca/events/2022/5-day-workshops/22w5055/videos/watch/202205051559-Lengerich.html)
 
 ## Contact Us

contextualized/easy/wrappers/SKLearnWrapper.py

@@ -30,6 +30,9 @@ def __init__(
         self.default_train_batch_size = 1
         self.default_test_batch_size = 16
         self.default_val_split = 0.2
+        self.default_encoder_width = 25
+        self.default_encoder_layers = 3
+        self.default_encoder_link_fn = LINK_FUNCTIONS["identity"]
         self.n_bootstraps = 1
         self.models = None
         self.trainers = None
@@ -69,6 +72,11 @@ def __init__(
             "fit": [],
             "wrapper": [
                 "n_bootstraps",
+                "es_patience",
+                "es_monitor",
+                "es_mode",
+                "es_min_delta",
+                "es_verbose",
             ],
         }
         self._update_acceptable_kwargs("model", extra_model_kwargs)
@@ -96,7 +104,7 @@ def __init__(
             "layers", self.constructor_kwargs["encoder_kwargs"]["layers"]
         )
         self.constructor_kwargs["encoder_kwargs"]["link_fn"] = kwargs.get(
-            "encoder_link_fn", self.constructor_kwargs["encoder_kwargs"]["link_fn"]
+            "encoder_link_fn", self.constructor_kwargs["encoder_kwargs"].get("link_fn", self.default_encoder_link_fn)
         )
         self.not_constructor_kwargs = {
             k: v
@@ -165,6 +173,7 @@ def maybe_add_kwarg(category, kwarg, default_val):
                 )
             ],
         )
+        print(organized_kwargs["trainer"]["callback_constructors"])
         organized_kwargs["trainer"]["callback_constructors"].append(
             lambda i: ModelCheckpoint(
                 monitor=kwargs.get("es_monitor", "val_loss"),
@@ -260,13 +269,13 @@ def _build_dataloaders(self, model, train_data, val_data, **kwargs):
         :param **kwargs:
         """
         train_dataloader = self._build_dataloader(
-            model, kwargs.get("train_batch_size", 1), *train_data
+            model, kwargs.get("train_batch_size", self.default_train_batch_size), *train_data
         )
         if val_data is None:
             val_dataloader = None
         else:
             val_dataloader = self._build_dataloader(
-                model, kwargs.get("val_batch_size", 16), *val_data
+                model, kwargs.get("val_batch_size", self.default_val_batch_size), *val_data
             )
 
         return train_dataloader, val_dataloader
@@ -287,7 +296,9 @@ def maybe_add_constructor_kwarg(kwarg, default_val):
         maybe_add_constructor_kwarg("loss_fn", LOSSES["mse"])
         maybe_add_constructor_kwarg(
             "encoder_kwargs",
-            {"width": 25, "layers": 2, "link_fn": LINK_FUNCTIONS["identity"]},
+            {"width": kwargs.get("encoder_width", self.default_encoder_width),
+            "layers": kwargs.get("encoder_layers", self.default_encoder_layers),
+            "link_fn": kwargs.get("encoder_link_fn", self.default_encoder_link_fn)},
         )
         if kwargs.get("subtype_probabilities", False):
             constructor_kwargs["encoder_kwargs"]["link_fn"] = LINK_FUNCTIONS["softmax"]
@@ -308,11 +319,9 @@ def maybe_add_constructor_kwarg(kwarg, default_val):
 
     def predict(self, C, X, individual_preds=False, **kwargs):
         """
-
         :param C:
         :param X:
         :param individual_preds:  (Default value = False)
-
         """
         if not hasattr(self, "models") or self.models is None:
             raise ValueError(
@@ -336,10 +345,8 @@ def predict_params(
         self, C, individual_preds=False, model_includes_mus=True, **kwargs
     ):
         """
-
         :param C:
         :param individual_preds:  (Default value = False)
-
         """
         # Returns betas, mus
         if kwargs.get("uses_y", True):
@@ -363,15 +370,14 @@ def predict_params(
             else:
                 return np.mean(betas, axis=0), np.mean(mus, axis=0)
         betas = np.array(preds)
-        if individual_preds:
+        if not individual_preds:
             return np.mean(betas, axis=0)
         return betas
 
     def fit(self, *args, **kwargs):
         """
         :param *args: C, X, Y (optional)
         :param **kwargs:
-
         """
         self.models = []
         self.trainers = []