Remove skip_init tutorial (#3562)

Fixes [T237454734](https://www.internalfb.com/intern/tasks/?t=237454734)

skip_init was originally added to support skipping initialization during
load_state_dict, see pytorch/pytorch#29523

This has been superseded by model creation under the meta device context
manager and then loading with assign=True, see
https://docs.pytorch.org/tutorials/recipes/recipes/module_load_state_dict_tips.html#using-torch-device-meta

So this tutorial can be removed

---------

Co-authored-by: Svetlana Karslioglu <[email protected]>

Author: mikaylagawarecki

redirects.py

@@ -38,4 +38,5 @@
     "recipes/recipes_index.html": "../recipes_index.html",
     "recipes/torchserve_vertexai_tutorial.html": "../index.html",
     "unstable_source/vulkan_workflow.rst": "../index.html",
+    "unstable/skip_param_init.html": "https://docs.pytorch.org/tutorials/recipes/recipes/module_load_state_dict_tips.html",
 }

unstable_index.rst

@@ -45,15 +45,6 @@ decide if we want to upgrade the level of commitment or to fail fast.
    :link: unstable/semi_structured_sparse.html
    :tags: Model-Optimiziation
 
-.. Modules
-
-.. customcarditem::
-   :header: Skipping Module Parameter Initialization in PyTorch 1.10
-   :card_description: Describes skipping parameter initialization during module construction in PyTorch 1.10, avoiding wasted computation.
-   :image: _static/img/thumbnails/cropped/generic-pytorch-logo.png
-   :link: unstable/skip_param_init.html
-   :tags: Modules
-
 .. vmap
 
 .. customcarditem::