Fix warning in GPT-Neo example and enable use_cache

[lekurile]

This PR fixes the following warning in the test-gpt-neo.py text-generation example by explicitly specifying the max_length argument:

UserWarning: Neither `max_length` nor `max_new_tokens` have been set, `max_length` will default to 50 (`self.config.max_length`). Controlling `max_length` via the config is deprecated and `max_length` will be removed from the config in v5 of Transformers -- we recommend using `max_new_tokens` to control the maximum length of the generation. 

It also sets use_cache=True when calling the generator to stay consistent with other text-generation examples.