AttributeError: 'GO1ModelConfig' object has no attribute 'initializer_range'

[Yunchen0930]

how to solve this issue?
[WARNING|modeling_utils.py:2376] 2025-12-03 14:41:42,429 >> InternLM2ForCausalLM has generative capabilities, as prepare_inputs_for_generation is explicitly defined. However, it doesn't directly inherit from GenerationMixin. 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.
  [WARNING|modeling_utils.py:2376] 2025-12-03 14:41:42,429 >> InternLM2ForCausalLMGO1 has generative capabilities, as prepare_inputs_for_generation is explicitly defined. However, it doesn't directly inherit from GenerationMixin. 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.
  [INFO|modeling_utils.py:2341] 2025-12-03 14:41:44,784 >> Instantiating ActionExpertModel model under default dtype torch.bfloat16.
  [rank0]: Traceback (most recent call last):
  [rank0]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 486, in
  [rank0]: main(
  [rank0]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 408, in main
  [rank0]: tokenizer, model = build_go1_model(dataset_args, model_args, training_args, space_args)
  [rank0]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 283, in build_go1_model
  [rank0]: model = GO1Model.from_pretrained(
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 277, in _wrapper
  [rank0]: return func(*args, **kwargs)
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5048, in from_pretrained
  [rank0]: ) = cls._load_pretrained_model(
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5362, in _load_pretrained_model
  [rank0]: model._initialize_missing_keys(missing_keys + mismatched_keys, is_quantized)
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5894, in _initialize_missing_keys
  [rank0]: self.initialize_weights()
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
  [rank0]: return func(*args, **kwargs)
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2984, in initialize_weights
  [rank0]: self.smart_apply(self._initialize_weights)
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2975, in smart_apply
  [rank0]: module.smart_apply(module._initialize_weights)
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2978, in smart_apply
  [rank0]: fn(self)
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2952, in _initialize_weights
  [rank0]: self._init_weights(module)
  [rank0]: File "/data2/yc/AgiBot-World/go1/internvl/model/go1/modeling_action_expert.py", line 513, in _init_weights
  [rank0]: std = self.config.initializer_range
  [rank0]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/configuration_utils.py", line 207, in getattribute
  [rank0]: return super().getattribute(key)
  [rank0]: AttributeError: 'GO1ModelConfig' object has no attribute 'initializer_range'
  [rank0]:[W1203 14:41:45.627508439 ProcessGroupNCCL.cpp:1496] Warning: WARNING: destroy_process_group() was not called before program exit, which can leak resources. For more info, please see https://pytorch.org/docs/stable/distributed.html#shutdown (function operator())
  [rank3]: Traceback (most recent call last):
  [rank3]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 486, in
  [rank3]: main(
  [rank3]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 408, in main
  [rank3]: tokenizer, model = build_go1_model(dataset_args, model_args, training_args, space_args)
  [rank3]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 283, in build_go1_model
  [rank3]: model = GO1Model.from_pretrained(
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 277, in _wrapper
  [rank3]: return func(*args, **kwargs)
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5048, in from_pretrained
  [rank3]: ) = cls._load_pretrained_model(
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5362, in _load_pretrained_model
  [rank3]: model._initialize_missing_keys(missing_keys + mismatched_keys, is_quantized)
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5894, in _initialize_missing_keys
  [rank3]: self.initialize_weights()
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
  [rank3]: return func(*args, **kwargs)
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2984, in initialize_weights
  [rank3]: self.smart_apply(self._initialize_weights)
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2975, in smart_apply
  [rank3]: module.smart_apply(module._initialize_weights)
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2978, in smart_apply
  [rank3]: fn(self)
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2952, in _initialize_weights
  [rank3]: self._init_weights(module)
  [rank3]: File "/data2/yc/AgiBot-World/go1/internvl/model/go1/modeling_action_expert.py", line 513, in _init_weights
  [rank3]: std = self.config.initializer_range
  [rank3]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/configuration_utils.py", line 207, in getattribute
  [rank3]: return super().getattribute(key)
  [rank3]: AttributeError: 'GO1ModelConfig' object has no attribute 'initializer_range'
  [rank2]: Traceback (most recent call last):
  [rank2]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 486, in
  [rank2]: main(
  [rank2]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 408, in main
  [rank2]: tokenizer, model = build_go1_model(dataset_args, model_args, training_args, space_args)
  [rank2]: File "/data2/yc/AgiBot-World/go1/internvl/train/go1_train.py", line 283, in build_go1_model
  [rank2]: model = GO1Model.from_pretrained(
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 277, in _wrapper
  [rank2]: return func(*args, **kwargs)
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5048, in from_pretrained
  [rank2]: ) = cls._load_pretrained_model(
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5362, in _load_pretrained_model
  [rank2]: model._initialize_missing_keys(missing_keys + mismatched_keys, is_quantized)
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 5894, in _initialize_missing_keys
  [rank2]: self.initialize_weights()
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 116, in decorate_context
  [rank2]: return func(*args, **kwargs)
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2984, in initialize_weights
  [rank2]: self.smart_apply(self._initialize_weights)
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2975, in smart_apply
  [rank2]: module.smart_apply(module._initialize_weights)
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2978, in smart_apply
  [rank2]: fn(self)
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/modeling_utils.py", line 2952, in _initialize_weights
  [rank2]: self._init_weights(module)
  [rank2]: File "/data2/yc/AgiBot-World/go1/internvl/model/go1/modeling_action_expert.py", line 513, in _init_weights
  [rank2]: std = self.config.initializer_range
  [rank2]: File "/home/hwuser/anaconda3/envs/go1/lib/python3.10/site-packages/transformers/configuration_utils.py", line 207, in getattribute
  [rank2]: return super().getattribute(key)
  [rank2]: AttributeError: 'GO1ModelConfig' object has no attribute 'initializer_range'

[ModiShi]

Which pretrained ckpt did you use? Is it from our open-sourced ckpt on HuggingFace?

[Yunchen0930]

Which pretrained ckpt did you use? Is it from our open-sourced ckpt on HuggingFace?

Yes,it is from your open-sourced ckpt on HuggingFace.
https://huggingface.co/agibot-world/GO-1
I added two lines of code and that solved the problem.

[ModiShi]

Thanks for the update, could you share which two lines you added? We haven't encountered this issue and it seems the action expert weights didn't load properly.

[lia0791]

Thanks for the update, could you share which two lines you added? We haven't encountered this issue and it seems the action expert weights didn't load properly.
I met same problem, and I add code in deploy.py/class GO1Infer:
if not hasattr(self.config, "initializer_range"):
self.config.initializer_range = 0.02
The problem has been temporarily solved

[wadeKeith]

PR Submitted ✨

Fixed in PR #156: #156

The fix adds initializer_range = 0.02 as a default attribute in GO1ModelConfig.__init__. This resolves the AttributeError that occurred when loading GO-1 pretrained checkpoints from HuggingFace.