Fixed missing import, added type annotation for IDE suggestions

+ Added type annotation for WaveRNN and Tacotron in train and
  generate files for each model
* Fixed missing import for numpy in `fatchord_version.py` and
  `deepmind_version.py`

Author: Ryan Butler

gen_wavernn.py

@@ -7,7 +7,7 @@
 import argparse
 
 
-def gen_testset(model, test_set, samples, batched, target, overlap, save_path):
+def gen_testset(model: WaveRNN, test_set, samples, batched, target, overlap, save_path):
 
     k = model.get_step() // 1000
 
@@ -34,7 +34,7 @@ def gen_testset(model, test_set, samples, batched, target, overlap, save_path):
         _ = model.generate(m, save_str, batched, target, overlap, hp.mu_law)
 
 
-def gen_from_file(model, load_path, save_path, batched, target, overlap):
+def gen_from_file(model: WaveRNN, load_path, save_path, batched, target, overlap):
 
     k = model.get_step() // 1000
     file_name = load_path.split('/')[-1]

models/deepmind_version.py

@@ -3,6 +3,7 @@
 import torch.nn.functional as F
 from utils.display import *
 from utils.dsp import *
+import numpy as np
 
 class WaveRNN(nn.Module):
     def __init__(self, hidden_size=896, quantisation=256):

models/fatchord_version.py

@@ -5,6 +5,7 @@
 from utils.display import *
 from utils.dsp import *
 import os
+import numpy as np
 
 
 class ResBlock(nn.Module):

train_tacotron.py

@@ -15,7 +15,7 @@
 def np_now(x): return x.detach().cpu().numpy()
 
 
-def tts_train_loop(model, optimizer, train_set, lr, train_steps, attn_example):
+def tts_train_loop(model: Tacotron, optimizer, train_set, lr, train_steps, attn_example):
     device = next(model.parameters()).device  # use same device as model parameters
 
     for p in optimizer.param_groups: p['lr'] = lr
@@ -81,7 +81,7 @@ def tts_train_loop(model, optimizer, train_set, lr, train_steps, attn_example):
         print(' ')
 
 
-def create_gta_features(model, train_set, save_path):
+def create_gta_features(model: Tacotron, train_set, save_path):
     device = next(model.parameters()).device  # use same device as model parameters
 
     iters = len(train_set)

train_wavernn.py

@@ -15,7 +15,7 @@
 import os
 
 
-def voc_train_loop(model, loss_func, optimizer, train_set, test_set, lr, total_steps):
+def voc_train_loop(model: WaveRNN, loss_func, optimizer, train_set, test_set, lr, total_steps):
     # Use same device as model parameters
     device = next(model.parameters()).device
 
@@ -56,6 +56,7 @@ def voc_train_loop(model, loss_func, optimizer, train_set, test_set, lr, total_s
                 if np.isnan(grad_norm):
                     print('grad_norm was NaN!')
             optimizer.step()
+            
             running_loss += loss.item()
 
             speed = i / (time.time() - start)