LoRA support in vllm and hf backends

[vicky-xef]

By default, there is no requirement to load LoRA adapter.

In vllm.py file:

• Added lora_request attribute (initialized to None, meaning LoRA is not used). This value is passed to every engine.generate(), engine.add_request() call.
• Added set_lora() method that loads a LoRA adapter by updating the lora_request attribute.
• Added clear_lora() method that resets lora_request to None (disabling LoRA usage).
• The user should set enable_lora=True in the engine_opts.

In hf.py file:

• Added a load_lora() method that loads a LoRA adapter.
• Added a set_lora() method that activates the loaded LoRA adapter.
• Added a clear_lora() method that deactivates the LoRA adapter.

[codecov]

Codecov Report

❌ Patch coverage is 15.78947% with 16 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
genlm/backend/llm/hf.py	25.00%	9 Missing ⚠️
genlm/backend/llm/vllm.py	0.00%	7 Missing ⚠️

📢 Thoughts on this report? Let us know!