[ESM] add accepts_loss_kwargs=False to EsmPreTrainedModel (#41006)

add accepts_loss_kwargs=False to EsmPreTrainedModel

Signed-off-by: Peter St. John <[email protected]>
Co-authored-by: Marc Sun <[email protected]>

Author: pstjohn

src/transformers/models/esm/modeling_esm.py

@@ -589,6 +589,7 @@ class EsmPreTrainedModel(PreTrainedModel):
     config: EsmConfig
     base_model_prefix = "esm"
     supports_gradient_checkpointing = True
+    accepts_loss_kwargs = False
     _no_split_modules = ["EsmLayer", "EsmFoldTriangularSelfAttentionBlock", "EsmEmbeddings"]
     _keys_to_ignore_on_load_unexpected = ["position_embeddings.weight"]
     _supports_flash_attn = True