fix autodocstring

src/lerobot/policies/pi0_openpi/transformers_replace/models/gemma/modeling_gemma.py

@@ -45,6 +45,16 @@ from .configuration_gemma import GemmaConfig
 logger = logging.get_logger(__name__)
 
 
+# Workaround for Python 3.10+ UnionType compatibility with transformers auto_docstring
+def safe_auto_docstring(func):
+    """Auto docstring decorator that handles Python 3.10+ UnionType gracefully."""
+    try:
+        return auto_docstring(func)
+    except (AttributeError, TypeError):
+        # If auto_docstring fails due to UnionType, just return the function unchanged
+        return func
+
+
 class GemmaRMSNorm(nn.Module):
     def __init__(self, dim: int, eps: float = 1e-6, cond_dim: int | None = None):
         super().__init__()
@@ -391,7 +401,7 @@ class GemmaDecoderLayer(GradientCheckpointingLayer):
         return outputs
 
 
-@auto_docstring
+@safe_auto_docstring
 class GemmaPreTrainedModel(PreTrainedModel):
     config_class = GemmaConfig
     base_model_prefix = "model"
@@ -422,7 +432,7 @@ class GemmaPreTrainedModel(PreTrainedModel):
                 module.weight.data.fill_(1.0)
 
 
-@auto_docstring
+@safe_auto_docstring
 class GemmaModel(GemmaPreTrainedModel):
     def __init__(self, config: GemmaConfig):
         super().__init__(config)
@@ -449,7 +459,7 @@ class GemmaModel(GemmaPreTrainedModel):
         self.embed_tokens = value
 
     @can_return_tuple
-    @auto_docstring
+    @safe_auto_docstring
     def forward(
         self,
         input_ids: torch.LongTensor | None = None,
@@ -567,7 +577,7 @@ class GemmaModel(GemmaPreTrainedModel):
 class KwargsForCausalLM(FlashAttentionKwargs, LossKwargs): ...
 
 
-@auto_docstring
+@safe_auto_docstring
 class GemmaForCausalLM(GemmaPreTrainedModel, GenerationMixin):
     _tied_weights_keys = ["lm_head.weight"]
     _tp_plan = {"lm_head": "colwise_rep"}
@@ -601,7 +611,7 @@ class GemmaForCausalLM(GemmaPreTrainedModel, GenerationMixin):
         return self.model
 
     @can_return_tuple
-    @auto_docstring
+    @safe_auto_docstring
     def forward(
         self,
         input_ids: torch.LongTensor | None = None,
@@ -685,7 +695,7 @@ class GemmaForCausalLM(GemmaPreTrainedModel, GenerationMixin):
         )
 
 
-@auto_docstring(
+@safe_auto_docstring(
     custom_intro="""
     The Gemma Model transformer with a sequence classification head on top (linear layer).
 
@@ -716,7 +726,7 @@ class GemmaForSequenceClassification(GemmaPreTrainedModel):
         self.model.embed_tokens = value
 
     @can_return_tuple
-    @auto_docstring
+    @safe_auto_docstring
     def forward(
         self,
         input_ids: torch.LongTensor | None = None,
@@ -792,7 +802,7 @@ class GemmaForSequenceClassification(GemmaPreTrainedModel):
         )
 
 
-@auto_docstring
+@safe_auto_docstring
 class GemmaForTokenClassification(GemmaPreTrainedModel):
     def __init__(self, config):
         super().__init__(config)
@@ -817,7 +827,7 @@ class GemmaForTokenClassification(GemmaPreTrainedModel):
         self.model.embed_tokens = value
 
     @can_return_tuple
-    @auto_docstring
+    @safe_auto_docstring
     def forward(
         self,
         input_ids: torch.LongTensor | None = None,