[Bugfix][Pytorch] Fix model save and load bug of stgcn_wave (#3303)

Co-authored-by: Jinjing Zhou <VoVAllen@users.noreply.github.com>

examples/pytorch/stgcn_wave/README.md

@@ -3,12 +3,12 @@ Spatio-Temporal Graph Convolutional Networks
 
 - Paper link: [arXiv](https://arxiv.org/pdf/1709.04875v4.pdf)
 - Author's code repo: https://github.com/VeritasYin/STGCN_IJCAI-18.
-Dependencies
-------------
-- PyTorch 1.1.0+
-- sklearn
-- dgl
-- tables
+- See [this blog](https://towardsdatascience.com/build-your-first-graph-neural-network-model-to-predict-traffic-speed-in-20-minutes-b593f8f838e5) for more details about running the code.
+- Dependencies
+  - PyTorch 1.1.0+
+  - sklearn
+  - dgl
+  - tables
 
 
 How to run

examples/pytorch/stgcn_wave/main.py

@@ -116,7 +116,7 @@ for epoch in range(1, epochs + 1):
     print("epoch", epoch, ", train loss:", l_sum / n, ", validation loss:", val_loss)
 
     
-best_model = STGCN_WAVE(blocks, n_his, n_route, G, drop_prob, num_layers).to(device)
+best_model = STGCN_WAVE(blocks, n_his, n_route, G, drop_prob, num_layers, device, args.control_str).to(device)
 best_model.load_state_dict(torch.load(save_path))
 
 

examples/pytorch/stgcn_wave/model.py

@@ -75,7 +75,7 @@ class STGCN_WAVE(nn.Module):
         super(STGCN_WAVE, self).__init__()
         self.control_str = control_str # model structure controller
         self.num_layers = len(control_str)
-        self.layers = []
+        self.layers = nn.ModuleList([])
         cnt = 0
         diapower = 0
         for i in range(self.num_layers):