Fix : Small fixes in DPO trainer args in DPO notebook

[ash-01xor]

Changes Made

Small fixes in the parameters present in the DPO trainer. While training the SmolLM instruct model on the "trl-lib/ultrafeedback_binarized dataset , using the following arguments which were present

beta=0.1,
# Maximum length of the input prompt in tokens
max_prompt_length=1024,
# Maximum combined length of prompt + response in tokens
max_length=1536

resulted in unexpected keyword argument errors.

Felt it would be better if the user can modify it based on their need and dataset used , rather than these arguments present by default and resulting in errors.

[ash-01xor]

@burtenshaw can you take a look at this