Fix step_11 hint descendents->descendants

[walkerbrown]

• Fixed hint in step_11.py and in docs: descendents -> descendants
• Fixed typo "You've completed built a complete GPT-2 implementation from scratch."
• Added hints to top of step_11.py regarding HuggingFace transformers and defaults()

# TODO: Import required modules and components from prior steps
# Hint: You'll need defaults() from max.tensor

# TODO: Import GPT2LMHeadModel and GPT2Tokenizer HuggingFace transformers library
# Hint: from transformers import GPT2LMHeadModel, GPT2Tokenizer