cascade: remove dead weight init code (Comfy-Org#13026)

This weight init process is fully shadowed be the weight load and
doesnt work in dynamic_vram were the weight allocation is deferred.

Author: rattus128

comfy/ldm/cascade/stage_a.py

@@ -136,16 +136,7 @@ def __init__(self, c, c_hidden):
             ops.Linear(c_hidden, c),
         )
 
-        self.gammas = nn.Parameter(torch.zeros(6), requires_grad=True)
-
-        # Init weights
-        def _basic_init(module):
-            if isinstance(module, nn.Linear) or isinstance(module, nn.Conv2d):
-                torch.nn.init.xavier_uniform_(module.weight)
-                if module.bias is not None:
-                    nn.init.constant_(module.bias, 0)
-
-        self.apply(_basic_init)
+        self.gammas = nn.Parameter(torch.zeros(6), requires_grad=False)
 
     def _norm(self, x, norm):
         return norm(x.permute(0, 2, 3, 1)).permute(0, 3, 1, 2)