[Minor] peft bug fix: HF peft version and tokenizer path in peft scripts

[realAsma]

1. HF peft LoraLayer init API have changed v0.7.0. With HF peft >= 0.7.0, I get the error in the attached error while trying to run examples/peft/peft_lora_clm_instruction_tuning.py:

    return GPTQLoraLinear(
  File "/workspace/AutoGPTQ/auto_gptq/utils/peft_utils.py", line 61, in __init__
    torch.nn.Linear.__init__(self, in_features, out_features)
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/torch/nn/modules/linear.py", line 96, in __init__
    self.weight = Parameter(torch.empty((out_features, in_features), **factory_kwargs))
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1712, in __setattr__
    self.register_parameter(name, value)
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/torch/nn/modules/module.py", line 577, in register_parameter
    elif hasattr(self, name) and name not in self._parameters:
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/peft/tuners/tuners_utils.py", line 358, in weight
    weight = base_layer.weight
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/peft/tuners/tuners_utils.py", line 358, in weight
    weight = base_layer.weight
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/peft/tuners/tuners_utils.py", line 358, in weight
    weight = base_layer.weight
  [Previous line repeated 975 more times]
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/peft/tuners/tuners_utils.py", line 352, in weight
    base_layer = self.get_base_layer()
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/peft/tuners/tuners_utils.py", line 341, in get_base_layer
    while hasattr(base_layer, "base_layer"):
  File "/home/akuriparambi/anaconda3/envs/agptq/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1695, in __getattr__
    raise AttributeError(f"'{type(self).__name__}' object has no attribute '{name}'")
RecursionError: maximum recursion depth exceeded while calling a Python object

I have added the left installation instructions to the examples README.

2. A minor bug in examples/peft - tokenizer_name_or_path is used to initialize AutoTokenizer instead of model_name_or_path:
   model_name_or_path points to the GPTQ saved model which might not have the tokenizer saved. tokenizer_name_or_path should be used instead.

[YooSungHyun]

what version do you use auto_gptq?
on my side, error raised can't set attribute 'active_adapter'

i used peft 0.6.2 and auto-gptq 0.6.0+cu118