[fix]: correct condition for restoring layer embeddings in TransformerDecorator/AdaptiveLayerLoss

[emapco]

This pull request makes a small fix to the logic that restores the original output settings in the call_grow_cache method of AdaptiveLayerLoss.py. The change ensures that the all_layer_embeddings key is deleted from the output only if the original setting for output_hidden_states was False.

• Fixed conditional logic in call_grow_cache to correctly remove all_layer_embeddings from the output only when output_hidden_states was originally False (sentence_transformers/losses/AdaptiveLayerLoss.py).

[Copilot]

Pull Request Overview

This PR fixes a logic error in the AdaptiveLayerLoss class where the condition for removing all_layer_embeddings from the output was inverted.

• Corrects the conditional logic to properly clean up all_layer_embeddings when it was not originally requested

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 1 comment.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

The call_use_cache method returns cached data without calling original_forward, so the input features parameter will not contain all_layer_embeddings. This check will always fail because features is the raw input to the forward method, not the output. If all_layer_embeddings needs to be included when output_hidden_states is enabled, it should be reconstructed from the cached self.embeddings data instead of reading from features.

Suggested change

	if self.transformer.auto_model.config.output_hidden_states and "all_layer_embeddings" in features:
	output["all_layer_embeddings"] = features["all_layer_embeddings"]
	if self.transformer.auto_model.config.output_hidden_states:
	# Reconstruct all_layer_embeddings from cached self.embeddings
	output["all_layer_embeddings"] = self.embeddings[self.call_idx]

[emapco]

self.embeddings doesn't contain all of the hidden states/layer embeddings. Perhaps call_grow_cache() should also storeself.all_layer_embeddings