fix(hf_ptq): announce the forced --skip_generate, and state its real reason

The flag flips silently, so a user who expected the post-quantization sample just
sees it missing; every other consequence of export_dir announces itself. The
comment also predated the tail-mutation fix -- finalize() converts the
non-decoder modules in place on every run, not only a resumed one, which is how
model_calib already words it.

Signed-off-by: Fridah-nv <201670829+Fridah-nv@users.noreply.github.com>

Author: Fridah-nv

examples/hf_ptq/hf_ptq.py

@@ -1237,7 +1237,10 @@ def _layerwise_get(cfg, key, default=None):
                 "layerwise.export_dir is not supported with an AutoQuantize recipe; "
                 "use a PTQ recipe, or drop export_dir and export afterwards."
             )
-        # A resumed run leaves skipped layers without amax; only the checkpoint is whole.
+        # finalize() converts the non-decoder modules in place without rolling them back, so
+        # the model is never valid for inference; a resumed run also skips layers' amax.
+        if not args.skip_generate:
+            print("Layerwise export: forcing --skip_generate, the model is left in export form.")
         args.skip_generate = True
 
     if args.batch_size == 0: