训练完internvl2-8b后无法merge lora

[H-cool]

可以正常训练，但是训练后无法使用如下命令进行merge lora.
CUDA_VISIBLE_DEVICES=0 swift export
--ckpt_dir '/root/hao/ocr/result/id_internvl_0106_v18/v2-20250106-164114/checkpoint-1500' --merge_lora true
报错如下：
From 👉v4.50👈 onwards, PreTrainedModel will NOT inherit from GenerationMixin, and this model will lose the ability to call generate and other related functions.

• If you're using trust_remote_code=True, you can get rid of this warning by loading the model with an auto class. See https://huggingface.co/docs/transformers/en/model_doc/auto#auto-classes
• If you are the owner of the model architecture code, please modify your model class such that it inherits from GenerationMixin (after PreTrainedModel, otherwise you'll get an exception).
• If you are not the owner of the model architecture class, please contact the model code owner to update it.
  Loading checkpoint shards: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████| 4/4 [00:05<00:00, 1.36s/it]
  [INFO:swift] default_system: 你是由上海人工智能实验室联合商汤科技开发的书生多模态大模型，英文名叫InternVL, 是一个有用无害的人工智能助手。
  [INFO:swift] Merge LoRA...
  [INFO:swift] Saving merged weights...
  [2025-01-07 15:25:24,141] [INFO] [real_accelerator.py:222:get_accelerator] Setting ds_accelerator to cuda (auto detect)
  Traceback (most recent call last):
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/swift/cli/export.py", line 5, in
  export_main()
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/swift/llm/export/export.py", line 41, in export_main
  return SwiftExport(args).main()
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/swift/llm/base.py", line 45, in main
  result = self.run()
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/swift/llm/export/export.py", line 24, in run
  merge_lora(args)
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/swift/llm/export/merge_lora.py", line 42, in merge_lora
  save_checkpoint(
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/swift/llm/utils.py", line 206, in save_checkpoint
  model.save_pretrained(output_dir, safe_serialization=safe_serialization, max_shard_size=max_shard_size)
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2815, in save_pretrained
  misplaced_generation_parameters = model_to_save.config._get_non_default_generation_parameters()
  File "/root/miniforge3/envs/hao_swift/lib/python3.10/site-packages/transformers/configuration_utils.py", line 1063, in _get_non_default_generation_parameters
  default_config = self.class()
  File "/root/.cache/huggingface/modules/transformers_modules/internvl2_8b/configuration_internvl_chat.py", line 50, in init
  if llm_config.get('architectures')[0] == 'LlamaForCausalLM':
  TypeError: 'NoneType' object is not subscriptable

[Jintao-Huang]

use transformers==4.44.*