Add api info for torch._C._nn.pyi (#148405)

APis involved are as followed:

- adaptive_avg_pool2d
- adaptive_avg_pool3d
- binary_cross_entropy
- col2im

ISSUE Related:
https://github.com/pytorch/pytorch/issues/148404
Pull Request resolved: https://github.com/pytorch/pytorch/pull/148405
Approved by: https://github.com/ezyang

tools/pyi/gen_pyi.py

@@ -412,6 +412,16 @@ def gen_nn_functional(fm: FileManager) -> None:
                         "tuple[Tensor, Tensor]",
                     )
                 ],
+                f"adaptive_avg_pool{d}d": [
+                    defs(
+                        f"adaptive_avg_pool{d}d",
+                        [
+                            INPUT,
+                            "output_size: _int | _size",
+                        ],
+                        "Tensor",
+                    )
+                ],
             }
         )
 
@@ -516,6 +526,31 @@ def gen_nn_functional(fm: FileManager) -> None:
                     "Tensor",
                 )
             ],
+            "binary_cross_entropy": [
+                defs(
+                    "binary_cross_entropy",
+                    [
+                        INPUT,
+                        "target: Tensor",
+                        "weight: Tensor | None = None",
+                        "reduction: str = ...",
+                    ],
+                    "Tensor",
+                )
+            ],
+            "col2im": [
+                defs(
+                    "col2im",
+                    [
+                        INPUT,
+                        "output_size: _int | _size",
+                        KERNEL_SIZE,
+                        "dilation: _int | _size",
+                        *STRIDE_PADDING,
+                    ],
+                    "Tensor",
+                )
+            ],
         }
     )