[Bug]: 'DeepSeekMultiTokenPredictor' object has no attribute 'embed_tokens'

[makabaka6338]

Your current environment

vllm serve /model/DeepSeek-R1 --host 0.0.0.0 --port 8000 --max-num-seqs 1024 --tensor-parallel-size 8 --speculative-config '{"method": "deepseek_mtp", "num_speculative_tokens": 8}'

🐛 Describe the bug

vllm v0.9.1执行报错
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] WorkerProc failed to start.
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] Traceback (most recent call last):
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/executor/multiproc_executor.py", line 466, in worker_main
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] worker = WorkerProc(*args, **kwargs)
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] ^^^^^^^^^^^^^^^^^^^^^^^^^^^
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/executor/multiproc_executor.py", line 363, in init
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] self.worker.load_model()
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_worker.py", line 180, in load_model
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] self.model_runner.load_model()
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/worker/gpu_model_runner.py", line 1618, in load_model
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] self.drafter.load_model(self.model)
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] File "/usr/local/lib/python3.12/dist-packages/vllm/v1/spec_decode/eagle.py", line 334, in load_model
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] and self.model.model.embed_tokens.weight.shape
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] File "/usr/local/lib/python3.12/dist-packages/torch/nn/modules/module.py", line 1940, in getattr
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] raise AttributeError(
(VllmWorker rank=5 pid=2573) ERROR 07-01 19:03:19 [multiproc_executor.py:492] AttributeError: 'DeepSeekMultiTokenPredictor' object has no attribute 'embed_tokens'
Loading safetensors checkpoint shards: 80% Completed | 130/163 [00:45<00:11, 2.87it/s]
(VllmWorker rank=0 pid=2568)

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[lengrongfu]

you can use latest version try it, lately having about pr #20022.