Change BasicNorm by adding 1+eps denominator; fix to (unused) DoubleSwish, revert to old status.

egs/librispeech/ASR/pruned_transducer_stateless7/scaling.py

@@ -491,8 +491,10 @@ class BasicNorm(torch.nn.Module):
             # gradients to allow the parameter to get back into the allowed
             # region if it happens to exit it.
             eps = eps.clamp(min=self.eps_min, max=self.eps_max)
+        eps = eps.exp()
         scales = (
-            torch.mean(x ** 2, dim=self.channel_dim, keepdim=True) + eps.exp()
+            (torch.mean(x ** 2, dim=self.channel_dim, keepdim=True) + eps) /
+            (1.0 + eps)
         ) ** -0.5
         return x * scales
 
@@ -1330,9 +1332,8 @@ class MaxEig(torch.nn.Module):
 
 class DoubleSwishFunction(torch.autograd.Function):
     """
-      double_swish(x) = x * (torch.sigmoid(x-1) + alpha)
+      double_swish(x) = x * torch.sigmoid(x-1)
 
-    for e.g. alpha=-0.05 (user supplied).
     This is a definition, originally motivated by its close numerical
     similarity to swish(swish(x)), where swish(x) =  x * sigmoid(x).
 
@@ -1356,18 +1357,8 @@ class DoubleSwishFunction(torch.autograd.Function):
         s = torch.sigmoid(x - 1.0)
         y = x * s
 
-        alpha = -0.05
-        beta = 0.05
-        x_limit = 0.15
-
-        # another part of this formula is:
-        # ... + 0.2 * x.clamp(min=-0.15, max=0.15)
-        # the deriv of this is
-        # beta * (x.abs() < x_limit).
-
         if requires_grad:
-            deriv =  (y * (1 - s) + s)  # ignores the alpha part.
-            deriv = deriv + (x.abs() < x_limit) * beta
+            deriv =  (y * (1 - s) + s)
 
             # notes on derivative of x * sigmoid(x - 1):
             # https://www.wolframalpha.com/input?i=d%2Fdx+%28x+*+sigmoid%28x-1%29%29
@@ -1376,7 +1367,7 @@ class DoubleSwishFunction(torch.autograd.Function):
             # the combination of "+ torch.rand_like(deriv)" and casting to torch.uint8 (which
             # floors), should be expectation-preserving.
             floor = -0.044
-            ceil = 1.2 + beta
+            ceil = 1.2
             d_scaled = ((deriv - floor) * (255.0 / (ceil - floor)) + torch.rand_like(deriv))
             if __name__ == "__main__":
                 # for self-testing only.
@@ -1384,8 +1375,6 @@ class DoubleSwishFunction(torch.autograd.Function):
                 assert d_scaled.max() < 256.0
             d_int = d_scaled.to(torch.uint8)
             ctx.save_for_backward(d_int)
-#  on wolframalpha, do: (x * sigmoid(x-1) - 0.05 * x  + 0.05 * min(0.15, max(-0.15, x)) + 0.025)  from x=-3 to 2
-        y = y + alpha * x + beta * x.clamp(min=-x_limit, max=x_limit) - 0.025
         if x.dtype == torch.float16 or torch.is_autocast_enabled():
             y = y.to(torch.float16)
         return y
@@ -1394,13 +1383,11 @@ class DoubleSwishFunction(torch.autograd.Function):
     def backward(ctx, y_grad: Tensor) -> Tensor:
         d, = ctx.saved_tensors
         # the same constants as used in forward pass.
-        alpha = -0.05
-        beta = 0.05
         floor = -0.043637
-        ceil = 1.2 + beta
+        ceil = 1.2
 
         d = (d * ((ceil - floor) / 255.0) + floor)
-        return (y_grad * (d + alpha))
+        return y_grad * d
 
 class DoubleSwish(torch.nn.Module):
     def __init__(self):
@@ -1412,7 +1399,7 @@ class DoubleSwish(torch.nn.Module):
         that we approximate closely with x * sigmoid(x-1).
         """
         if torch.jit.is_scripting():
-            return x * (torch.sigmoid(x - 1.0) - 0.05) + 0.05 * x.clamp(min=-0.15, max=0.15)
+            return x * torch.sigmoid(x - 1.0)
         return DoubleSwishFunction.apply(x)
 
 
@@ -1741,8 +1728,6 @@ def _test_basic_norm():
     y_rms = (y ** 2).mean().sqrt()
     print("x rms = ", x_rms)
     print("y rms = ", y_rms)
-    assert y_rms < x_rms
-    assert y_rms > 0.5 * x_rms
 
 
 def _test_double_swish_deriv():