[BUG fix] Rebase caused spec decode fix

[xuechendi]

Error reported in https://jira.habana-labs.com/browse/SW-212516

Found two recent merged PR breaks down Spec Decode functionality:

1. Support mllama (llama 3.2) model for HPU #491 overrides existing workerwrapperBase design for speculative decoding.

if model_runner_cls is not None:
    ModelRunnerClass = model_runner_cls

is not needed since we now use codes as below for init model_runner_cls to follow upstream design.

if model_runner_cls is not None:
            self.model_runner = model_runner_cls(self.model_runner)

2. Prepare sin/cos buffers for rope outside model forward #566 is not working in Spec Decode Eagle mode
   Due to input tensors is now different to the pre-assumption that decode_fwd only provide one token per seq. Spec Decode provides multiple candidates tokens as q.
   To fix that, added a new ENV - "VLLM_COS_SIN_RECOMPUTE=true", need to use it to trigger recompute to cos and sin for spec decode.

[xuechendi]

@michalkuligowski , please help to review.

[xuechendi]

@kzawora-intel , please check a fix here:
previous mllama PR will break spec decode, I added a fix PR
de79b5c