Support model saving and loading (#316)

cornac/eval_methods/cross_validation.py

@@ -128,13 +128,17 @@ def _next_fold(self):
 
     def evaluate(self, model, metrics, user_based, show_validation):
         result = CVResult(model.name)
+
         for _ in range(self.n_folds):
             self._get_train_test()
+            new_model = model.clone()  # clone a completely new model
             fold_result, _ = BaseMethod.evaluate(
-                self, model, metrics, user_based, show_validation=False
+                self, new_model, metrics, user_based, show_validation=False
             )
             result.append(fold_result)
             self._next_fold()
+
         result.organize()
+
         return result, None  # no validation result of CV
 

cornac/experiment/experiment.py

@@ -13,6 +13,9 @@
 # limitations under the License.
 # ============================================================================
 
+import os
+from datetime import datetime
+
 from .result import ExperimentResult
 from .result import CVExperimentResult
 from ..metrics.rating import RatingMetric
@@ -42,6 +45,10 @@ class Experiment:
 
     show_validation: bool, optional, default: True 
         Whether to show the results on validation set (if exists).
+        
+    save_dir: str, optional, default: None
+        Path to a directory for storing trained models and logs. If None, 
+        models will NOT be stored and logs will be saved in the current working directory.
 
     Attributes
     ----------
@@ -63,13 +70,15 @@ def __init__(
         user_based=True,
         show_validation=True,
         verbose=False,
+        save_dir=None,
     ):
         self.eval_method = eval_method
         self.models = self._validate_models(models)
         self.metrics = self._validate_metrics(metrics)
         self.user_based = user_based
         self.show_validation = show_validation
         self.verbose = verbose
+        self.save_dir = save_dir
         self.result = None
         self.val_result = None
 
@@ -110,6 +119,7 @@ def _create_result(self):
                 self.val_result = ExperimentResult()
 
     def run(self):
+        """Run the Cornac experiment"""
         self._create_result()
 
         for model in self.models:
@@ -119,11 +129,23 @@ def run(self):
                 user_based=self.user_based,
                 show_validation=self.show_validation,
             )
+
             self.result.append(test_result)
             if self.val_result is not None:
                 self.val_result.append(val_result)
 
+            if not isinstance(self.result, CVExperimentResult):
+                model.save(self.save_dir)
+
+        output = ""
         if self.val_result is not None:
-            print("\nVALIDATION:\n...\n{}".format(self.val_result))
+            output += "\nVALIDATION:\n...\n{}".format(self.val_result)
+        output += "\nTEST:\n...\n{}".format(self.result)
+
+        print(output)
 
-        print("\nTEST:\n...\n{}".format(self.result))
+        timestamp = datetime.now().strftime("%Y-%m-%d_%H-%M-%S-%f")
+        save_dir = "." if self.save_dir is None else self.save_dir
+        output_file = os.path.join(save_dir, "CornacExp-{}.log".format(timestamp))
+        with open(output_file, "w") as f:
+            f.write(output)

cornac/models/baseline_only/recom_bo.pyx

@@ -15,6 +15,8 @@
 
 # cython: language_level=3
 
+import multiprocessing
+
 cimport cython
 from cython.parallel import prange
 from cython cimport floating, integral
@@ -23,8 +25,10 @@ from libc.math cimport abs
 
 import numpy as np
 cimport numpy as np
+from tqdm import trange
 
 from ..recommender import Recommender
+from ...utils.init_utils import zeros
 
 
 class BaselineOnly(Recommender):
@@ -74,17 +78,28 @@ class BaselineOnly(Recommender):
         self.learning_rate = learning_rate
         self.lambda_reg = lambda_reg
         self.early_stop = early_stop
-        self.init_params = {} if init_params is None else init_params
         self.seed = seed
 
-        import multiprocessing
         if seed is not None:
             self.num_threads = 1
         elif num_threads > 0 and num_threads < multiprocessing.cpu_count():
             self.num_threads = num_threads
         else:
             self.num_threads = multiprocessing.cpu_count()
 
+        # Init params if provided
+        self.init_params = {} if init_params is None else init_params
+        self.u_biases = self.init_params.get('Bu', None)
+        self.i_biases = self.init_params.get('Bi', None)
+        self.global_mean = 0.0
+
+    def _init(self):
+        n_users, n_items = self.train_set.num_users, self.train_set.num_items
+
+        self.global_mean = self.train_set.global_mean
+        self.u_biases = zeros(n_users) if self.u_biases is None else self.u_biases
+        self.i_biases = zeros(n_items) if self.i_biases is None else self.i_biases
+
     def fit(self, train_set, val_set=None):
         """Fit the model to observations.
 
@@ -102,14 +117,9 @@ class BaselineOnly(Recommender):
         """
         Recommender.fit(self, train_set, val_set)
 
-        n_users, n_items = train_set.num_users, train_set.num_items
-        self.global_mean = train_set.global_mean
-
-        from ...utils.init_utils import zeros
-        self.u_biases = self.init_params.get('Bu', zeros(n_users))
-        self.i_biases = self.init_params.get('Bi', zeros(n_items))
-
         if self.trainable:
+            self._init()
+
             (rid, cid, val) = train_set.uir_tuple
             self._fit_sgd(rid, cid, val.astype(np.float32), self.u_biases, self.i_biases)
 
@@ -139,7 +149,6 @@ class BaselineOnly(Recommender):
             floating r, r_pred, error, delta_loss
             integral u, i, j
 
-        from tqdm import trange
         progress = trange(max_iter, disable=not self.verbose)
         for epoch in progress:
             last_loss = loss