add multi lora support

[sven-knoblauch]

in addition to the last pull reqeust, now you can also use multiple lora adapters.

Now you can use a single adapter as input (as before): {"name": "xxx", "path": "xxx/xxxxx", "base_model_name": "xxx/xxxx"}

And also as a list of adapters: [{"name": "xxx", "path": "xxx/xxxxx", "base_model_name": "xxx/xxxx"},{"name": "xxx", "path": "xxx/xxxxx", "base_model_name": "xxx/xxxx"},...]

[sven-knoblauch]

Thanks @pandyamarut for merging the last lora adapter Pull Request. Just added a small update to include also multi lora support.

[nerdylive123]

Hey mind writng some simple examples to the readme.md too? would be nice if theres an example usage

[BeautyyuYanli]

Waiting for merge please. Seems the LoRA support is broken due to some conflicts since v1.9 https://github.com/runpod-workers/worker-vllm/blob/6fc770415def3b65ddf8c3a80a8b36fb1454f8e7/src/engine.py#L145-L151

[sven-knoblauch]

should be working now with the new version