Add LoRA-FA support

**Is your feature request related to a problem? Please describe:**
LoRA is designed to reduce the memory requirements to finetune LLMs and already exists in Fairseq2 to some capacity. [LoRA-FA](https://arxiv.org/abs/2308.03303) reduces the memory overhead of [vanilla LoRA](https://arxiv.org/abs/2106.09685) even further by freezing the A matrix while leaving the B matrix as trainable.

**Describe the solution you would like:**
Because LoRA-FA is effectively an extension of LoRA, there should be an additional parameter in the LoRAConfig class that is used to effectively enable/disable LoRa-FA.

**Describe the alternatives you have considered:**
None

**Additional Context:**
None