Add use_auth_token arg to sft_trainer example

[corey-lambda]

This is for huggingface models that require using an access token to download (like meta's llama2). Tested by running the following command:

> huggingface-cli login
> python trl/examples/scripts/sft_trainer.py \
    --model_name meta-llama/Llama-2-7b-hf \
    --dataset_name timdettmers/openassistant-guanaco \
    --load_in_4bit \
    --use_peft \
    --batch_size 4 \
    --gradient_accumulation_steps 2 \
    --use_auth_token

[lvwerra]

Thanks a lot for the fix! Made a small change so this works out of the box without updating the existing examples.

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.