weightslab 1.2.4__tar.gz → 1.2.5__tar.gz

weightslab-1.2.5/CHANGELOG.md

@@ -0,0 +1 @@
+# Changelog - 2026-06-17 v1.2.5 (1)

{weightslab-1.2.4/weightslab.egg-info → weightslab-1.2.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: weightslab
-Version: 1.2.4
+Version: 1.2.5
 Summary: Paving the way between black-box and white-box modeling.
 Author-email: Alexandru-Andrei Rotaru <alexandru@graybx.com>, Guillaue Pelluet <guillaue@graybx.com>
 License: BSD 2-clause

{weightslab-1.2.4 → weightslab-1.2.5}/weightslab/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '1.2.4'
-__version_tuple__ = version_tuple = (1, 2, 4)
+__version__ = version = '1.2.5'
+__version_tuple__ = version_tuple = (1, 2, 5)
 
-__commit_id__ = commit_id = 'gf0148ea1e'
+__commit_id__ = commit_id = 'g935f5b0b4'

{weightslab-1.2.4 → weightslab-1.2.5}/weightslab/integrations/ultralytics/trainer.py

@@ -70,24 +70,26 @@ class WLAwareTrainer(DetectionTrainer):
             @wl.eval_fn
             def _validate(loader):
                 # Clean pause ctrl callbacks
+                import copy
                 raised_exc = None
+                val_loader = trainer.validator.dataloader
                 try:
-                    # Should be remove as guards is no blocker now
-                    on_val_batch_start_callbacks = trainer.validator.callbacks.pop('on_val_batch_start')  # Remove pause ctrl deps. We can use it as resume pause ctrl will trigger the training in //.
-                    on_val_batch_end_callbacks = trainer.validator.callbacks.pop('on_val_batch_end')  # Remove pause ctrl deps. We can use it as resume pause ctrl will trigger the training in //.
-                    val_loader = trainer.validator.dataloader
                     trainer.validator.dataloader = loader
-                    trainer.validator(model=(trainer.ema.ema if trainer.ema else trainer.model))
+                    # validator(model=...) takes UL's *standalone* path, which
+                    # wraps the model in AutoBackend and FUSES it in place
+                    # (conv+bn -> conv with bias, bn deleted). Fusing the live
+                    # EMA/model rewrites its state_dict keys, so the next
+                    # trainer.ema.update(model) raises KeyError 'model.*.conv.bias'.
+                    # Validate on a throwaway deep copy so the live EMA / training
+                    # model are never mutated. (`underlying` is the unwrapped model
+                    # captured above; avoids deep-copying the WL wrapper.)
+                    src = trainer.ema.ema if trainer.ema else underlying
+                    eval_model = copy.deepcopy(src).eval()
+                    trainer.validator(model=eval_model)
                 except Exception as e:
                     raised_exc = e
-                    pass
-
-                # Set Val loader
-                trainer.validator.dataloader = val_loader  # Reset val trainer
-
-                # Set val callbacks
-                trainer.validator.callbacks["on_val_batch_start"].extend(on_val_batch_start_callbacks)
-                trainer.validator.callbacks["on_val_batch_end"].extend(on_val_batch_end_callbacks)
+                finally:
+                    trainer.validator.dataloader = val_loader  # Reset val loader
 
                 # Finally raise exc.
                 if raised_exc is not None:

{weightslab-1.2.4 → weightslab-1.2.5/weightslab.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: weightslab
-Version: 1.2.4
+Version: 1.2.5
 Summary: Paving the way between black-box and white-box modeling.
 Author-email: Alexandru-Andrei Rotaru <alexandru@graybx.com>, Guillaue Pelluet <guillaue@graybx.com>
 License: BSD 2-clause

weightslab-1.2.4/CHANGELOG.md

@@ -1 +0,0 @@
-# Changelog - 2026-06-17 v1.2.4 (1)