add missing attribute _supports_param_buffer_assignment for gpt-j. (#32359)

Co-authored-by: Guoming Zhang <37257613+nv-guomingz@users.noreply.github.com>

src/transformers/models/gptj/modeling_gptj.py

@@ -500,6 +500,7 @@ class GPTJPreTrainedModel(PreTrainedModel):
     _no_split_modules = ["GPTJBlock"]
     _skip_keys_device_placement = "past_key_values"
     _supports_flash_attn_2 = True
+    _supports_param_buffer_assignment = False
 
     def __init__(self, *inputs, **kwargs):
         super().__init__(*inputs, **kwargs)