chore : correct typos [no ci] (#20041)

* fix(docs): correct typos found during code review

Non-functional changes only:
- Fixed minor spelling mistakes in comments
- Corrected typos in user-facing strings
- No variables, logic, or functional code was modified.

Signed-off-by: Marcel Petrick <mail@marcelpetrick.it>

* Update docs/backend/CANN.md

Co-authored-by: Aaron Teo <taronaeo@gmail.com>

* Revert "Auxiliary commit to revert individual files from 846d1c301281178efbc6ce6060ad34c1ebe45af8"

This reverts commit 02fcf0c7db661d5ff3eff96b2b2db9fdb7213256.

* Update tests/test-backend-ops.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

* Update tests/test-backend-ops.cpp

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

---------

Signed-off-by: Marcel Petrick <mail@marcelpetrick.it>
Co-authored-by: Aaron Teo <taronaeo@gmail.com>
Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>

ggml/src/ggml-metal/ggml-metal-device.m

@@ -1281,7 +1281,7 @@ struct ggml_metal_buffer {
     bool use_residency_sets;
 
     // optional MTLResidencySet
-    // note: cannot use explicity "id<MTLResidencySet>" here because it is not available on certain OSes
+    // note: cannot use explicitly "id<MTLResidencySet>" here because it is not available on certain OSes
     id rset;
 
     // pointers to global device

ggml/src/ggml-metal/ggml-metal-ops.cpp

@@ -631,7 +631,7 @@ int ggml_metal_op_acc(ggml_metal_op_t ctx, int idx) {
     const bool inplace = (bool) ((const int32_t *) op->op_params)[4];
 
     if (!inplace) {
-        // run a separete kernel to cpy src->dst
+        // run a separate kernel to cpy src->dst
         // not sure how to avoid this
         // TODO: make a simpler cpy_bytes kernel
 
@@ -1644,7 +1644,7 @@ int ggml_metal_op_set(ggml_metal_op_t ctx, int idx) {
     const bool inplace = (bool) ((const int32_t *) op->op_params)[4];
 
     if (!inplace) {
-        // run a separete kernel to cpy src->dst
+        // run a separate kernel to cpy src->dst
         // not sure how to avoid this
         // TODO: make a simpler cpy_bytes kernel
 
@@ -2005,7 +2005,7 @@ int ggml_metal_op_mul_mat(ggml_metal_op_t ctx, int idx) {
         const int16_t r0ptg  = nypsg*nsg;         // num src0 rows per threadgroup
               int16_t r1ptg  = 4;                 // num src1 rows per threadgroup
 
-        // note: not sure how optimal are those across all different hardware. there might be someting cleverer
+        // note: not sure how optimal are those across all different hardware. there might be something cleverer
         switch (ne11) {
             case 2:
                 r1ptg = 2; break;

ggml/src/ggml-metal/ggml-metal.cpp

@@ -14,7 +14,7 @@
 #define GGML_METAL_MAX_DEVICES 16
 
 // number of Metal devices
-// note: can be overriden with GGML_METAL_DEVICES env to simulate virtual devices
+// note: can be overridden with GGML_METAL_DEVICES env to simulate virtual devices
 static int g_devices = 1;
 
 ////////////////////////////////////////////////////////////////////////////////

ggml/src/ggml-metal/ggml-metal.metal

@@ -4218,7 +4218,7 @@ kernel void kernel_im2col(
 template [[host_name("kernel_im2col_f32")]] kernel im2col_t kernel_im2col<float>;
 template [[host_name("kernel_im2col_f16")]] kernel im2col_t kernel_im2col<half>;
 
-// TODO: obolete -- remove
+// TODO: obsolete -- remove
 //typedef void (im2col_ext_t)(
 //        constant ggml_metal_kargs_im2col & args,
 //        device const float * x,