add custom RMSNorm to ALL_LAYERNORM_LAYERS (huggingface#26227)

* add LlamaRMSNorm to ALL_LAYERNORM_LAYERS

* fixup

* add IdeficsRMSNorm to ALL_LAYERNORM_LAYERS and fixup

src/transformers/models/idefics/modeling_idefics.py

@@ -31,6 +31,7 @@
 from ...activations import ACT2FN
 from ...modeling_outputs import ModelOutput
 from ...modeling_utils import PretrainedConfig
+from ...pytorch_utils import ALL_LAYERNORM_LAYERS
 from ...utils import (
     add_start_docstrings,
     add_start_docstrings_to_model_forward,
@@ -261,7 +262,7 @@ def freeze_model(model, module_exceptions=[]):
     }
     module_exceptions_mapped = [mapping[m] for m in module_exceptions]
     for module in model.modules():
-        if module_exceptions and any([isinstance(module, t) for t in module_exceptions_mapped]):
+        if module_exceptions and any(isinstance(module, t) for t in module_exceptions_mapped):
             module.requires_grad_(True)  # Explicitely setting it to true to avoid any mistakes
         else:
             module.requires_grad_(False)
@@ -496,6 +497,9 @@ def forward(self, hidden_states):
         return self.weight * hidden_states
 
 
+ALL_LAYERNORM_LAYERS.append(IdeficsRMSNorm)
+
+
 # this was adapted from LlamaRotaryEmbedding
 class IdeficsEmbedding(torch.nn.Module):
     def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):

src/transformers/models/llama/modeling_llama.py

@@ -30,6 +30,7 @@
 from ...activations import ACT2FN
 from ...modeling_outputs import BaseModelOutputWithPast, CausalLMOutputWithPast, SequenceClassifierOutputWithPast
 from ...modeling_utils import PreTrainedModel
+from ...pytorch_utils import ALL_LAYERNORM_LAYERS
 from ...utils import add_start_docstrings, add_start_docstrings_to_model_forward, logging, replace_return_docstrings
 from .configuration_llama import LlamaConfig
 
@@ -89,6 +90,9 @@ def forward(self, hidden_states):
         return self.weight * hidden_states.to(input_dtype)
 
 
+ALL_LAYERNORM_LAYERS.append(LlamaRMSNorm)
+
+
 class LlamaRotaryEmbedding(nn.Module):
     def __init__(self, dim, max_position_embeddings=2048, base=10000, device=None):
         super().__init__()