Fix static_quantization_tutorial error in qat_model

The model must be in eval mode to fuse and in train mode to
prepare_qat. After this fix, the tutorial runs to completion on a
linux x86 system.

Fixes #1269

Signed-off-by: BJ Hargrave <[email protected]>

Author: bjhargrave

Diff for: advanced_source/static_quantization_tutorial.rst

@@ -533,7 +533,9 @@ We fuse modules as before
 .. code:: python
 
     qat_model = load_model(saved_model_dir + float_model_file)  
+    qat_model.eval()
     qat_model.fuse_model()  
+    qat_model.train()
 
     optimizer = torch.optim.SGD(qat_model.parameters(), lr = 0.0001) 
     # The old 'fbgemm' is still available but 'x86' is the recommended default.