From b362f16dcd487aae4c47f5d30fbc4b3f78f748b3 Mon Sep 17 00:00:00 2001
From: Vincent Jeanselme <vincent.jeanselme@gmail.com>
Date: Thu, 31 Dec 2020 18:48:16 +0000
Subject: [PATCH 1/2] Evaluation

Compute the loss on a given set of points
---
 dsm/dsm_api.py | 14 +++++++++++++-
 1 file changed, 13 insertions(+), 1 deletion(-)

diff --git a/dsm/dsm_api.py b/dsm/dsm_api.py
index 6988728..028f6b2 100644
--- a/dsm/dsm_api.py
+++ b/dsm/dsm_api.py
@@ -31,7 +31,7 @@
 from dsm.dsm_torch import DeepRecurrentSurvivalMachinesTorch
 from dsm.losses import predict_cdf
 import dsm.losses as losses
-from dsm.utilities import train_dsm, _get_padded_features, _get_padded_targets
+from dsm.utilities import train_dsm, _get_padded_features, _get_padded_targets, _reshape_tensor_with_nans
 
 import torch
 import numpy as np
@@ -118,6 +118,18 @@ def fit(self, x, t, e, vsize=0.15,
     self.fitted = True
     
     return self
+  
+  def eval(self, x, t, e):
+    """
+        Evaluates models with conditional loss
+    """
+    processed_data = self._prepocess_training_data(x, t, e, 0, 0)
+    _, _, _, x_val, t_val, e_val = processed_data
+    x_val, t_val, e_val = x_val,\
+        _reshape_tensor_with_nans(t_val),\
+        _reshape_tensor_with_nans(e_val)
+    return losses.conditional_loss(self.torch_model, x_val, t_val, e_val,
+                                  elbo=False).detach().numpy()
 
   def _prepocess_test_data(self, x):
     return torch.from_numpy(x)

From 4b9b090761adde2e8aef95a05eea8ff6b537a706 Mon Sep 17 00:00:00 2001
From: Vincent Jeanselme <vincent.jeanselme@gmail.com>
Date: Sat, 2 Jan 2021 15:36:57 +0000
Subject: [PATCH 2/2] Update function name and import

---
 dsm/dsm_api.py | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/dsm/dsm_api.py b/dsm/dsm_api.py
index 028f6b2..3e69dca 100644
--- a/dsm/dsm_api.py
+++ b/dsm/dsm_api.py
@@ -31,7 +31,8 @@
 from dsm.dsm_torch import DeepRecurrentSurvivalMachinesTorch
 from dsm.losses import predict_cdf
 import dsm.losses as losses
-from dsm.utilities import train_dsm, _get_padded_features, _get_padded_targets, _reshape_tensor_with_nans
+from dsm.utilities import train_dsm, _get_padded_features, _get_padded_targets,\
+  _reshape_tensor_with_nans
 
 import torch
 import numpy as np
@@ -119,7 +120,7 @@ def fit(self, x, t, e, vsize=0.15,
     
     return self
   
-  def eval(self, x, t, e):
+  def _eval_conditional_loss(self, x, t, e):
     """
         Evaluates models with conditional loss
     """