Fix the pad_token_id error

[yuanwu2017]

1. Fix the following error:

File "/mnt/disk4/yuanwu/workspace/trl/examples/research_projects/stack_llama/scripts/rl_training.py", line 263, in <module>
  pipe_outputs = sentiment_pipe(texts, **sent_kwargs)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/pipelines/text_classification.py", line 156, in __call__
  result = super().__call__(*args, **kwargs)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/pipelines/base.py", line 1143, in __call__
  outputs = list(final_iterator)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/pipelines/pt_utils.py", line 124, in __next__
  item = next(self.iterator)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/pipelines/pt_utils.py", line 125, in __next__
  processed = self.infer(item, **self.params)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/pipelines/base.py", line 1068, in forward
  model_outputs = self._forward(model_inputs, **forward_params)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/pipelines/text_classification.py", line 187, in _forward
  return self.model(**model_inputs)
File "/mnt/disk4/yuanwu/env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
  return self._call_impl(*args, **kwargs)
File "/mnt/disk4/yuanwu/env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
  return forward_call(*args, **kwargs)
File "/mnt/disk4/yuanwu/workspace/accelerate/src/accelerate/hooks.py", line 166, in new_forward
  output = module._old_forward(*args, **kwargs)
File "/mnt/disk4/yuanwu/workspace/transformers/src/transformers/models/llama/modeling_llama.py", line 1369, in forward
  raise ValueError("Cannot handle batch sizes > 1 if no padding token is defined.")
ValueError: Cannot handle batch sizes > 1 if no padding token is defined.

2. Add the load_in_8bit arguments for rl_training script.

[younesbelkada]

Thank you for the fix @yuanwu2017 ! can you run the styling checks? make precommit

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[yuanwu2017]

Done. Thanks

[younesbelkada]

Thanks for your work on this!