From a1c9af2bad21d49ee5cc258308bafd3e04dae6cb Mon Sep 17 00:00:00 2001
From: Chintan Shah <chintan.shah@apprenticehealth.com>
Date: Mon, 30 Sep 2019 21:58:55 -0400
Subject: [PATCH] Setup curriculum learning framework

---
 model/pytorch/dcrnn_model.py | 15 +++++++++++++++
 1 file changed, 15 insertions(+)

diff --git a/model/pytorch/dcrnn_model.py b/model/pytorch/dcrnn_model.py
index 6b8b489..5612b1d 100644
--- a/model/pytorch/dcrnn_model.py
+++ b/model/pytorch/dcrnn_model.py
@@ -1,3 +1,4 @@
+import numpy as np
 import torch
 import torch.nn as nn
 from abc import ABC, abstractmethod
@@ -122,6 +123,20 @@ class DecoderModel(nn.Module, DCRNNModel):
                                                      bias=True) for _ in
                                           range(self.num_rnn_layers - 1)]
 
+    def t_step_forward_pass(self, hidden_state, inputs, output, t):
+        cell_input = inputs[:, t, :]  # (batch_size, input_size)
+
+        if self.is_training:
+            if t > 0 and self.use_curriculum_learning:
+                c = np.random.uniform(0, 1)
+                if c >= self._compute_sampling_threshold(): #todo
+                    cell_input = output[
+                        t - 1]  # todo: this won't work because the linear layer is applied after forward_impl
+
+        cell_output, hidden_state = self._forward_cell(cell_input, hidden_state)
+        output[t] = cell_output
+        return hidden_state
+
     def forward(self, inputs, hidden_state=None):
         """
         Decoder forward pass.