Friendli-model-optimizer

- Support ExaoneForCausalLM
- Fix CohereForCausalLM with better quantization logic.

- Further optimization for running FP8, and INT8 quantization.
- Support searching automatic calibration dataset batch size for running FMO.
- Support [AWQ(Activation-aware Weight Quantization)].

**Full Changelog**: https://github.com/friendliai/friendli-model-optimizer/compare/v0.2.1...v0.3.0

**Full Changelog**: https://github.com/friendliai/friendli-model-optimizer/compare/v0.2.0...v0.2.1