DSM library change

README.md

@@ -4,12 +4,13 @@ This repository is a Pytorch implementation of the paper :  [Dynamic Deep Hit](h
 
 The code is based on the [Deep Survival Machines repository](https://github.com/autonlab/DeepSurvivalMachines) and use a similar module structure.
 
-
 ### Code Description
 All the code related to Dynamic Deep Hit is loacted in `ddh`. Only API should be used for application.
 
 An example is available in the `examples` folder.
 
-## TODOs
-- Full documentation
-- Further testing to ensure the model is reproducing the original paper's performance
+## Note
+- Deep Survival Machines library has been merged into auton_survival.
+```
+pip install auton_survival 
+```

ddh/ddh_api.py

@@ -1,5 +1,5 @@
-from DeepSurvivalMachines.dsm.dsm_api import DeepRecurrentSurvivalMachines
-from dsm.utilities import _reshape_tensor_with_nans, _get_padded_features, _get_padded_targets
+from auton_survival.models.dsm import DeepRecurrentSurvivalMachines
+from auton_survival.models.dsm.utilities import _reshape_tensor_with_nans, _get_padded_features, _get_padded_targets
 from ddh.ddh_torch import DynamicDeepHitTorch
 from ddh.utils import train_ddh
 from ddh.losses import total_loss

ddh/utils.py

@@ -1,5 +1,5 @@
 from ddh.losses import total_loss
-from dsm.utilities import get_optimizer, _reshape_tensor_with_nans
+from auton_survival.models.dsm.utilities import get_optimizer, _reshape_tensor_with_nans
 
 import torch
 import torch.nn as nn