ValueError: not enough values to unpack (expected 4, got 2)

[Second222None]

Describe the bug

Since there is no powerful GPU, we have to run the End2End test with a samll model Qwen/Qwen2-1.5B. LMcache could start successfully. We get the error when sending requests to server.

The error from current screen:

And the error from /tmp/root-8000-stdout.log

(VllmWorkerProcess pid=73509) ERROR 11-20 22:14:12 multiproc_worker_utils.py:233]     _, _, num_heads, head_size = kv_cache[0].shape
(VllmWorkerProcess pid=73509) ERROR 11-20 22:14:12 multiproc_worker_utils.py:233] ValueError: not enough values to unpack (expected 4, got 2) 

env:

• OS: Ubuntu 22.04.5 LTS
• vllm: v0.6.2 (pip install vllm==0.6.2)
• LMcache: v0.1.3-alpha (installed from source)
• lmcache-vllm: v0.6.2.2 (installed from source)
• GPU: Tesla T4 (16GB) x 2

To Reproduce
Steps to reproduce the behavior:

1. Set to use Qwen/Qwen2-1.5B by changing tests/tests.py. def test_lmcache_local_cpu(model = "Qwen/Qwen2-1.5B") -> pd.DataFrame:
2. python3 main.py tests/tests.py -f test_lmcache_local_cpu -o outputs/