[WIP] Enable DeepSeek Models on Intel Gaudi device

[YangQun1]

Example

DeepSeek-V2-Lite

Run on single Gaudi2

python examples/runtime/engine/offline_batch_inference.py \
    --device hpu \
    --model-path deepseek-ai/DeepSeek-V2-Lite \
    --trust-remote-code \
    --disable-mla

python3 -m sglang.bench_one_batch \
    --batch-size 1 \
    --input 1024 \
    --output 8 \
    --model deepseek-ai/DeepSeek-V2-Lite \
    --trust-remote-code \
    --device hpu \
    --disable-mla

DeepSeek-R1

Run on x8 Gaudi3

python3 -m sglang.bench_one_batch \
    --batch-size 1 \
    --input 1024 \
    --output 8 \
    --model /software/data/DeepSeek-R1 \
    --trust-remote-code \
    --device hpu \
    --tp 8 \
    --load-format dummy \
    --disable-mla