Add missing arguments

Signed-off-by: cyy <cyyever@outlook.com>

Author: cyyever

src/transformers/models/d_fine/modular_d_fine.py

@@ -898,7 +898,7 @@ def __init__(self, config: DFineConfig):
 
 class DFineForObjectDetection(RTDetrForObjectDetection, DFinePreTrainedModel):
     def __init__(self, config: DFineConfig):
-        DFinePreTrainedModel.__init__(config)
+        DFinePreTrainedModel.__init__(self, config)
 
         # D-FINE encoder-decoder model
         self.eval_idx = config.eval_idx if config.eval_idx >= 0 else config.decoder_layers + config.eval_idx

src/transformers/models/data2vec/modular_data2vec_audio.py

@@ -183,7 +183,7 @@ def load_adapter(self):
 
 class Data2VecAudioModel(Data2VecAudioPreTrainedModel, Wav2Vec2Model):
     def __init__(self, config: Data2VecAudioConfig):
-        Data2VecAudioPreTrainedModel.__init__(config)
+        Data2VecAudioPreTrainedModel.__init__(self, config)
         self.config = config
         self.feature_extractor = Data2VecAudioFeatureEncoder(config)
         self.feature_projection = Data2VecAudioFeatureProjection(config)
@@ -215,7 +215,7 @@ def forward(self, **super_kwargs):
 
 class Data2VecAudioForCTC(Data2VecAudioPreTrainedModel, Wav2Vec2ForCTC):
     def __init__(self, config):
-        Data2VecAudioPreTrainedModel.__init__(config)
+        Data2VecAudioPreTrainedModel.__init__(self, config)
 
         self.data2vec_audio = Data2VecAudioModel(config)
         self.dropout = nn.Dropout(config.final_dropout)

src/transformers/models/eomt/modular_eomt.py

@@ -249,7 +249,7 @@ class EomtPatchEmbeddings(Dinov2PatchEmbeddings):
 
 class EomtEmbeddings(Dinov2Embeddings, nn.Module):
     def __init__(self, config: EomtConfig) -> None:
-        Dinov2Embeddings().__init__()
+        Dinov2Embeddings().__init__(self, config)
 
         self.config = config
         self.patch_size = config.patch_size

src/transformers/models/ernie4_5/modular_ernie4_5.py

@@ -84,7 +84,7 @@ def apply_rotary_pos_emb(q, k, cos, sin, position_ids=None, unsqueeze_dim=1):
 
 class Ernie4_5MLP(LlamaMLP):
     def __init__(self, config: Ernie4_5Config):
-        super().__init__()
+        super().__init__(config)
 
         self.gate_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=config.use_bias)
         self.up_proj = nn.Linear(self.hidden_size, self.intermediate_size, bias=config.use_bias)

src/transformers/models/ernie4_5_moe/modular_ernie4_5_moe.py

@@ -314,7 +314,7 @@ def forward(
 @auto_docstring
 class Ernie4_5_MoeForCausalLM(MixtralForCausalLM, Ernie4_5_MoePreTrainedModel):
     def __init__(self, config):
-        Ernie4_5_MoePreTrainedModel().__init__(config)
+        Ernie4_5_MoePreTrainedModel().__init__(self, config)
         self.model = Ernie4_5_MoeModel(config)
         self.vocab_size = config.vocab_size
         self.lm_head = nn.Linear(config.hidden_size, config.vocab_size, bias=config.use_bias)

src/transformers/models/esm/openfold_utils/residue_constants.py

@@ -541,7 +541,6 @@ def make_bond_key(atom1_name: str, atom2_name: str) -> str:
 
 # A compact atom encoding with 14 columns
 # pylint: disable=line-too-long
-# pylint: disable=bad-whitespace
 restype_name_to_atom14_names: dict[str, list[str]] = {
     "ALA": ["N", "CA", "C", "O", "CB", "", "", "", "", "", "", "", "", ""],
     "ARG": ["N", "CA", "C", "O", "CB", "CG", "CD", "NE", "CZ", "NH1", "NH2", "", "", ""],
@@ -566,7 +565,6 @@ def make_bond_key(atom1_name: str, atom2_name: str) -> str:
     "UNK": ["", "", "", "", "", "", "", "", "", "", "", "", "", ""],
 }
 # pylint: enable=line-too-long
-# pylint: enable=bad-whitespace
 
 
 # This is the standard residue order when coding AA type as a number.

src/transformers/models/evolla/modular_evolla.py

@@ -65,7 +65,7 @@
 
 class EvollaSaProtEmbeddings(EsmEmbeddings):
     def __init__(self, config):
-        super().__init__()
+        super().__init__(config)
         # remove the position_ids in EsmEmbeddings
         self.position_ids = None
 

src/transformers/models/falcon_h1/modular_falcon_h1.py

@@ -252,7 +252,7 @@ def forward(
 
 class FalconH1RMSNormGated(MambaRMSNormGated):
     def __init__(self, hidden_size, eps=1e-6, n_groups=1, norm_before_gate=True):
-        super().__init__()
+        super().__init__(hidden_size=hidden_size, eps=eps)
         self.weight = nn.Parameter(torch.ones(hidden_size))
         self.variance_epsilon = eps
         self.n_groups = n_groups
@@ -813,7 +813,7 @@ def forward(
 
 class FalconH1MLP(LlamaMLP):
     def __init__(self, config: FalconH1Config = None):
-        super().__init__()
+        super().__init__(config)
         self.gate_multiplier, self.down_multiplier = config.mlp_multipliers
 
     def forward(self, x):

src/transformers/models/falcon_mamba/modular_falcon_mamba.py

@@ -517,7 +517,7 @@ class FalconMambaCausalLMOutput(MambaCausalLMOutput):
 
 class FalconMambaModel(MambaModel, FalconMambaPreTrainedModel):
     def __init__(self, config):
-        FalconMambaPreTrainedModel.__init__(config)
+        FalconMambaPreTrainedModel.__init__(self, config)
 
         self.embeddings = nn.Embedding(config.vocab_size, config.hidden_size)
         self.layers = nn.ModuleList(

src/transformers/models/gemma3/modular_gemma3.py

@@ -382,8 +382,7 @@ def __init__(self, config: Gemma3TextConfig):
 
 
 class Gemma3RMSNorm(Gemma2RMSNorm):
-    def __init__(self, dim: int, eps: float = 1e-6):
-        super().__init__()
+    pass
 
 
 class Gemma3RotaryEmbedding(Gemma2RotaryEmbedding):
@@ -396,7 +395,7 @@ class Gemma3Attention(Gemma2Attention):
     def __init__(self, config: Gemma3TextConfig, layer_idx: int):
         self.is_sliding = config.layer_types[layer_idx] == "sliding_attention"
 
-        super().__init__()
+        super().__init__(config, layer_idx)
         self.sliding_window = config.sliding_window if self.is_sliding else None
 
         self.q_norm = Gemma3RMSNorm(dim=config.head_dim, eps=config.rms_norm_eps)