Zephyr: report the host reference label alongside the top-5 (pytorch#23353)

`gen_input.py` writes `MV2_HOST_TOP1` and `MV2_HOST_LABEL` into
`mv2_input.h`, but nothing reads them. The sample prints bare class IDs,
so a reader cannot interpret the output on target without an ImageNet
label table.

This prints the recorded host prediction after the top-5 and states
whether the device top-1 agrees:

```
Top-5 predictions:
  [1] class 258: 6.1895
  [2] class 259: 4.8418
  [3] class 261: 3.4442

host float32 reference: class 258 (Samoyed) -> device top-1 MATCHES
```

Guarded on both defines, so headers generated before pytorch#23300 still build.
No label table is added to the target; the single string already comes
from the generated header.

## Verification

Flashed and run on an Alif Ensemble E8 DevKit (Cortex-M55 HP at 400 MHz
+ Ethos-U55, 256 MAC) with Zephyr 4.4.0 and Zephyr SDK 1.0.1, using a
photograph prepared by `gen_input.py` and an INT8 model calibrated on
250 ImageNet photos. Inference completed in 28 ms and the device top-1
matched the float32 torchvision reference.

Flash cost is 160 bytes.

Authored with Claude.

Author: psiddh

zephyr/samples/mv2-ethosu/src/main.cpp

@@ -269,6 +269,17 @@ void print_top_k(const std::vector<EValue>& outputs) {
         top_indices[j],
         static_cast<double>(top_values[j]));
   }
+
+#if defined(MV2_HOST_TOP1) && defined(MV2_HOST_LABEL)
+  // gen_input.py records what float32 torchvision predicts for this image, so
+  // the class ID can be interpreted without an ImageNet label table on target.
+  ET_LOG(
+      Info,
+      "\nhost float32 reference: class %d (%s) -> device top-1 %s",
+      MV2_HOST_TOP1,
+      MV2_HOST_LABEL,
+      (top_indices[0] == MV2_HOST_TOP1) ? "MATCHES" : "DIFFERS");
+#endif
 }
 
 } // namespace