[Bug]: get_model_info return structure is not consistent.

[SmartManoj]

What happened?

import litellm
from pprint import pprint
max_tokens = litellm.get_model_info("gpt-4")
pprint(max_tokens)

Now it returns max_input_tokens and max_output_tokens but the docstring is not mentioned this.
Return the same for huggingface custom_llm_provider and update the docstring.

Relevant log output

{'input_cost_per_token': 3e-05,
 'litellm_provider': 'openai',
 'max_input_tokens': 8192,
 'max_output_tokens': 4096,
 'max_tokens': 4096,
 'mode': 'chat',
 'output_cost_per_token': 6e-05,
 'supported_openai_params': ['frequency_penalty',
                             'logit_bias',
                             'logprobs',
                             'top_logprobs',
                             'max_tokens',
                             'n',
                             'presence_penalty',
                             'seed',
                             'stop',
                             'stream',
                             'stream_options',
                             'temperature',
                             'top_p',
                             'tools',
                             'tool_choice',
                             'function_call',
                             'functions',
                             'max_retries',
                             'extra_headers',
                             'user'],
 'supports_function_calling': True}

Twitter / LinkedIn details

No response

[SmartManoj]

Diff b/w max_output_tokens and max_tokens?