Feat: Onbaord PlamoForCausalLM Architecture

QEfficient/base/pytorch_transforms.py

@@ -107,6 +107,9 @@ def apply(cls, model: nn.Module) -> Tuple[nn.Module, bool]:
             ):
                 for orig_method_name, mapped_method in repl_method_map.items():
                     setattr(module, orig_method_name, MethodType(mapped_method, module))
+                    # Handling the __init__ calls in the models
+                    if hasattr(module, "__qeff_init__"):
+                        module.__qeff_init__()
                     transformed = True
 
         return model, transformed

QEfficient/transformers/models/modeling_auto.py

@@ -1291,6 +1291,7 @@ class QEFFAutoModelForCausalLM(QEFFBaseModel):
         FP8DeQuantLinearToLinearTransform,
         CustomOpsTransform,
         KVCacheTransform,
+        KVCacheModuleMethodMapperTransform,
     ]
     _onnx_transforms = [FP16ClipTransform, SplitTensorsTransform]
 

QEfficient/transformers/models/plamo/__init__.py

@@ -0,0 +1,6 @@
+# -----------------------------------------------------------------------------
+#
+# Copyright (c) 2025 Qualcomm Innovation Center, Inc. All rights reserved.
+# SPDX-License-Identifier: BSD-3-Clause
+#
+# -----------------------------------------------------------------------------