[BE] fix typos in functorch/ and scripts/ (#156081)

Pull Request resolved: https://github.com/pytorch/pytorch/pull/156081
Approved by: https://github.com/albanD
ghstack dependencies: #156080

scripts/build_android.sh

@@ -157,7 +157,7 @@ if [ -n "${USE_VULKAN}" ]; then
   fi
 fi
 
-# Use-specified CMake arguments go last to allow overridding defaults
+# Use-specified CMake arguments go last to allow overriding defaults
 CMAKE_ARGS+=($@)
 
 # Patch pocketfft (as Android does not have aligned_alloc even if compiled with c++17

scripts/build_mobile.sh

@@ -80,7 +80,7 @@ if [ "${VERBOSE:-}" == '1' ]; then
   CMAKE_ARGS+=("-DCMAKE_VERBOSE_MAKEFILE=1")
 fi
 
-# Use-specified CMake arguments go last to allow overridding defaults
+# Use-specified CMake arguments go last to allow overriding defaults
 CMAKE_ARGS+=("$@")
 
 # Now, actually build the Android target.

scripts/jit/log_extract.py

@@ -95,7 +95,7 @@ def run():
         "--no-nnc-dynamic",
         dest="nnc_dynamic",
         action="store_false",
-        help="DONT't benchmark nnc with dynamic shapes",
+        help="don't benchmark nnc with dynamic shapes",
     )
     parser.set_defaults(nnc_dynamic=False)
 

scripts/release_notes/apply_categories.py

@@ -1,4 +1,4 @@
-# Quick scipt to apply categorized items to the
+# Quick script to apply categorized items to the
 # base commitlist . Useful if you are refactoring any code
 # but want to keep the previous data on categories
 

scripts/release_notes/classifier.py

@@ -156,9 +156,9 @@ class CommitClassifier(nn.Module):
         elif isinstance(most_likely_index, torch.Tensor):
             return [self.categories[i] for i in most_likely_index]
 
-    def get_most_likely_category_name(self, inpt):
+    def get_most_likely_category_name(self, input):
         # Input will be a dict with title and author keys
-        logits = self.forward(inpt)
+        logits = self.forward(input)
         most_likely_index = torch.argmax(logits, dim=1)
         return self.convert_index_to_category_name(most_likely_index)
 
@@ -264,9 +264,9 @@ def generate_batch(batch):
 
 
 def train_step(batch, model, optimizer, loss):
-    inpt, targets = batch
+    input, targets = batch
     optimizer.zero_grad()
-    output = model(inpt)
+    output = model(input)
     l = loss(output, targets)
     l.backward()
     optimizer.step()
@@ -275,8 +275,8 @@ def train_step(batch, model, optimizer, loss):
 
 @torch.no_grad()
 def eval_step(batch, model, loss):
-    inpt, targets = batch
-    output = model(inpt)
+    input, targets = batch
+    output = model(input)
     l = loss(output, targets)
     return l