[Frontend] Support for chat completions input in the tokenize endpoint

[sasha0552]

Related: #5054 (#5054 (comment))

Example:

$ curl -H "Authorization: Bearer token-abc123" -H "Content-Type: application/json" -d '{"model":"meta-llama/Meta-Llama-3-8B-Instruct","prompt":"test"}' http://localhost:8000/tokenize
{
  "count": 1,
  "max_model_len": 8192,
  "tokens": [1985],
}

$ curl -H "Authorization: Bearer token-abc123" -H "Content-Type: application/json" -d '{"model":"meta-llama/Meta-Llama-3-8B-Instruct","messages":[{"role":"user","content":"Hi"}]}' http://localhost:8000/tokenize
{
  "count": 11,
  "max_model_len": 8192,
  "tokens": [128000,128006,882,128007,271,13347,128009,128006,78191,128007,271]
}

$ curl -H "Authorization: Bearer token-abc123" -H "Content-Type: application/json" -d '{"model":"meta-llama/Meta-Llama-3-8B-Instruct","tokens":[128000,128006,882,128007,271,13347,128009,128006,78191,128007,271]}' http://localhost:8000/detokenize
{
  "prompt": "<|begin_of_text|><|start_header_id|>user<|end_header_id|>\n\nHi<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n"
}

[DarkLight1337]

To speed up the CI queue for #5905, I've cancelled the distributed tests for the latest CI run in this PR since they won't pass anyway until #5905 has been merged. Please merge main into your branch after that happens so that the CI can pass once again.

[DarkLight1337]

Hmm, I think we can move this to a new file (as in create a new subclass of OpenAIServing) since it's not really a part of OpenAI's Chat or Completions API. We can support both Chat and Completions input.

[sasha0552]

Just a reminder for myself (so I don't cancel the CI).

[sasha0552]

/ready

[DarkLight1337]

LGTM, thanks for extending this feature!