Rewrite existing links to custom ops gdocs with the landing page (#127423)

NB: these links will be live after the docs build happens, which is once
a day.

Test Plan:
- existing tests

Pull Request resolved: https://github.com/pytorch/pytorch/pull/127423
Approved by: https://github.com/jansel, https://github.com/williamwen42
ghstack dependencies: #127291, #127292, #127400

aten/src/ATen/core/MetaFallbackKernel.cpp

@@ -16,8 +16,8 @@ static void metaFallback(
       "fake impl or Meta kernel registered. You may have run into this message "
       "while using an operator with PT2 compilation APIs (torch.compile/torch.export); "
       "in order to use this operator with those APIs you'll need to add a fake impl. "
-      "Please see the following doc for next steps:  "
-      "https://docs.google.com/document/d/1_W62p8WJOQQUzPsJYa7s701JXt0qf2OfLub2sbkHOaU/edit");
+      "Please see the following for next steps:  "
+      "https://pytorch.org/docs/main/notes/custom_operators.html");
 }
 
 TORCH_LIBRARY_IMPL(_, Meta, m) {

c10/core/StorageImpl.cpp

@@ -18,7 +18,7 @@ void throwNullDataPtrError() {
       "If you're using torch.compile/export/fx, it is likely that we are erroneously "
       "tracing into a custom kernel. To fix this, please wrap the custom kernel into "
       "an opaque custom op. Please see the following for details: "
-      "https://docs.google.com/document/d/1W--T6wz8IY8fOI0Vm8BF44PdBgs283QvpelJZWieQWQ");
+      "https://pytorch.org/docs/main/notes/custom_operators.html");
 }
 
 // NOTE: [FakeTensor.data_ptr deprecation]

c10/core/TensorImpl.h

@@ -1580,7 +1580,7 @@ struct C10_API TensorImpl : public c10::intrusive_ptr_target {
         "If you're using torch.compile/export/fx, it is likely that we are erroneously "
         "tracing into a custom kernel. To fix this, please wrap the custom kernel into "
         "an opaque custom op. Please see the following for details: "
-        "https://docs.google.com/document/d/1W--T6wz8IY8fOI0Vm8BF44PdBgs283QvpelJZWieQWQ\n"
+        "https://pytorch.org/docs/main/notes/custom_operators.html\n"
         "If you're using Caffe2, Caffe2 uses a lazy allocation, so you will need to call "
         "mutable_data() or raw_mutable_data() to actually allocate memory.");
     // Caller does the type check.

torch/_dynamo/output_graph.py

@@ -1676,7 +1676,7 @@ err_epilogue = (
     "(and fall back to eager-mode PyTorch) on all ops "
     "that have do not have the 'pt2_compliant_tag'. "
     "Please see the following doc for how to mark this op as PT2 compliant "
-    "https://docs.google.com/document/d/1W--T6wz8IY8fOI0Vm8BF44PdBgs283QvpelJZWieQWQ"
+    "https://pytorch.org/docs/main/notes/custom_operators.html"
 )
 
 

torch/library.py

@@ -556,7 +556,7 @@ def register_fake(
     This API may be used as a decorator (see examples).
 
     For a detailed guide on custom ops, please see
-    https://docs.google.com/document/d/1W--T6wz8IY8fOI0Vm8BF44PdBgs283QvpelJZWieQWQ/edit
+    https://pytorch.org/docs/main/notes/custom_operators.html
 
     Examples:
         >>> import torch