[GitHub] [tvm] sjain58 opened a new pull request, #15708: [Relay] Simplify Conv->bias_add->mul->add to Conv->bias_add

["sjain58 (via GitHub)" <gi...@apache.org>]

sjain58 opened a new pull request, #15708:
URL: https://github.com/apache/tvm/pull/15708

   Simplify Conv->bias_add->mul->add to Conv->bias_add sequence if one of the inputs to Conv, bias_add, mul and add are constant scalars.
   
   def @main(%q1: Tensor[(1, 3, 224, 224), float32]) {
   %0 = nn.conv2d(%q1, meta[relay.Constant][0], padding=[3, 3, 3, 3], channels=64, kernel_size=[7, 7]);
   %1 = nn.bias_add(%0, meta[relay.Constant][1]);
   %2 = multiply(%1, meta[relay.Constant][2]);
   add(%2, meta[relay.Constant][3])
   }
   
   Replace with
   def @main(%q1: Tensor[(1, 3, 224, 224), float32]) {
   %0 = reshape(meta[relay.Constant][1], newshape=[64, 1, 1, 1]);
   %1 = multiply(meta[relay.Constant][0], %0);
   %2 = reshape(meta[relay.Constant][2], newshape=[64, 1, 1]);
   %3 = multiply(%2, meta[relay.Constant][1]);
   %4 = add(%3, meta[relay.Constant][3]);
   %5 = nn.conv2d(%q1, %1, padding=[3, 3, 3, 3], channels=64, kernel_size=[7, 7]);
   %6 = reshape(%4, newshape=[64]);
   nn.bias_add(%5, %6)
   }
   
   res[p,q,r,s] = ({SUM{i=[0,c-1], j=[0,kh-1], k=[0,kw-1]}(a[p,i,r+j,s+k] * W[q,i,j,k])} + b[q]) * c1[q] + c2[q]
   res[p,q,r,s] = {SUM{i=[0,c-1], j=[0,kh-1], k=[0,kw-1]}(a[p,i,r+j,s+k] * W[q,i,j,k])} * c1[q] + b[q]c1[q] + c2[q]
   res[p,q,r,s] = Conv2d(a, Wc1) + bias_add(b*c1+c2)
   
   In the above, %1, %3, %4 are constants and can be folded, so we're left with 2 ops, as opposed to the original 4 ops.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] sjain58 closed pull request #15708: [Relay] Simplify Conv->bias_add->mul->add to Conv->bias_add

["sjain58 (via GitHub)" <gi...@apache.org>]

sjain58 closed pull request #15708: [Relay] Simplify Conv->bias_add->mul->add to Conv->bias_add
URL: https://github.com/apache/tvm/pull/15708


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] sjain58 commented on pull request #15708: [Relay] Simplify Conv->bias_add->mul->add to Conv->bias_add

["sjain58 (via GitHub)" <gi...@apache.org>]

sjain58 commented on PR #15708:
URL: https://github.com/apache/tvm/pull/15708#issuecomment-1721103684

   > Have you tried `FoldScaleAxis`? Combined with `SimplifyExpr` I think it already does such optimization.
   
   Thanks @masahi for the suggestion! `FoldScaleAxis` combined with `SimplifyExpr` works for me when I enable O3 optimizations. 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org

[GitHub] [tvm] masahi commented on pull request #15708: [Relay] Simplify Conv->bias_add->mul->add to Conv->bias_add

["masahi (via GitHub)" <gi...@apache.org>]

masahi commented on PR #15708:
URL: https://github.com/apache/tvm/pull/15708#issuecomment-1716237070

   Have you tried `FoldScaleAxis`? Combined with `SimplifyExpr` I think it already does such optimization. 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@tvm.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org