model.dequantize() does not remove quantization_config from model.config which effects in error when loading the model

[konradkalita]

System Info

• transformers version: 4.46.0
• bitsandbytes version: 0.44.1
• Platform: Linux-5.10.0-32-cloud-amd64-x86_64-with-glibc2.31
• Python version: 3.10.14
• Huggingface_hub version: 0.25.1
• Safetensors version: 0.4.5
• Accelerate version: 1.0.0
• Accelerate config: not found
• PyTorch version (GPU?): 2.4.1+cu121 (True)
• Tensorflow version (GPU?): not installed (NA)
• Flax version (CPU?/GPU?/TPU?): not installed (NA)
• Jax version: not installed
• JaxLib version: not installed
• Using distributed or parallel set-up in script?:
• Using GPU in script?: yes
• GPU type: NVIDIA A100-SXM4-40GB

Who can help?

quantization: @SunMarc @MekkCyber

Information

• [ ] The official example scripts
• [X] My own modified scripts

Tasks

• [ ] An officially supported task in the examples folder (such as GLUE/SQuAD, ...)
• [ ] My own task or dataset (give details below)

Reproduction

from transformers import AutoModel, AutoTokenizer
from transformers import BitsAndBytesConfig

bnb_config = {
    "bnb_4bit_compute_dtype": "bfloat16",
    "bnb_4bit_quant_type": "nf4",
    "bnb_4bit_use_double_quant": True,
    "load_in_4bit": True
}

model = AutoModel.from_pretrained("mistralai/Mistral-7B-v0.1", quantization_config=bnb_config)
model = model.dequantize()
model.save_pretrained("mistral_test_dequantized")
model = AutoModel.from_pretrained("mistral_test_dequantized")

Stacktrace:

Traceback (most recent call last):
  File "./test_model.py", line 40, in <module>
    main()
  File "./test_model.py", line 37, in main
    model = AutoModel.from_pretrained("mistral_test_dequantized")
  File "./venv/lib/python3.10/site-packages/transformers/models/auto/auto_factory.py", line 564, in from_pretrained
    return model_class.from_pretrained(
  File "./venv/lib/python3.10/site-packages/transformers/modeling_utils.py", line 4226, in from_pretrained
    ) = cls._load_pretrained_model(
  File "./venv/lib/python3.10/site-packages/transformers/modeling_utils.py", line 4729, in _load_pretrained_model
    new_error_msgs, offload_index, state_dict_index = _load_state_dict_into_meta_model(
  File "./venv/lib/python3.10/site-packages/transformers/modeling_utils.py", line 994, in _load_state_dict_into_meta_model
    hf_quantizer.create_quantized_param(model, param, param_name, param_device, state_dict, unexpected_keys)
  File "./venv/lib/python3.10/site-packages/transformers/quantizers/quantizer_bnb_4bit.py", line 207, in create_quantized_param
    raise ValueError(
ValueError: Supplied state dict for layers.0.mlp.down_proj.weight does not contain `bitsandbytes__*` and possibly other `quantized_stats` components.

Expected behavior

Dequantized model can be saved using save_pretrained and properly loaded using from_pretrained

[SunMarc]

Thanks for the report. This happens because the config still have the quantization_config. Would you like to open a PR to fix it ? I think all you need to do is to add in the dequantize() method the following: del self.config.quantization_config