[Activation Quantization] Dynamic Per Token Support

[dsikka]

Summary

• Add a CompressedTensorsW8A8DynamicToken scheme to support dynamic-per token activation quantization
• Update config parsing to support updates made to the config.json / quantization config provided with the model
• Update config parsing logic to pull in functionality from compressed_tensors; add in compressed_tensors as a requirement
• Update/add in logic for llama layer mappings when dealing with the ignore list
• Update to use QuantizationArgs directly from `compressed_tensors

TODO:

• config naming issue between sparseml and vllm