[Feature]: Qwen2.5 model : ValueError: This model does not support the 'embedding' task. Supported tasks: {'generate'}

[tarikbeijing]

🚀 The feature, motivation and pitch

Using Qwen2.5 model : ValueError: This model does not support the 'embedding' task. Supported tasks: {'generate'}

reproduction :
python -m vllm.entrypoints.openai.api_server --served-model-name Qwen2.5-1.5B-Instruct --model Qwen/Qwen2.5-1.5B-Instruct --max_model_len 17000 --dtype float16 --trust_remote_code --host 127.0.0.1 --port 8080 --uvicorn-log-level debug --api-key gUaq3eYuYZjuBfejwH-lVrFAlTbi9g3vQnRZD4jBCYA --task "embedding"

Qwen2.5 model is recognized as "Qwen2ForCausalLM": ("qwen2", "Qwen2ForCausalLM")

Qwen2Config {
"_name_or_path": "Qwen/Qwen2.5-1.5B-Instruct",
"architectures": [
"Qwen2ForCausalLM"
],

but in the Embedding model section it is not registred.

in config file : vllm/vllm/model_executor/models/registry.py
_EMBEDDING_MODELS = {
# [Text-only]
"BertModel": ("bert", "BertEmbeddingModel"),
"Gemma2Model": ("gemma2", "Gemma2EmbeddingModel"),
"MistralModel": ("llama", "LlamaEmbeddingModel"),
"Qwen2ForRewardModel": ("qwen2_rm", "Qwen2ForRewardModel"),
"Qwen2ForSequenceClassification": (
"qwen2_cls", "Qwen2ForSequenceClassification"),
# [Multimodal]
"LlavaNextForConditionalGeneration": ("llava_next", "LlavaNextForConditionalGeneration"), # noqa: E501
"Phi3VForCausalLM": ("phi3v", "Phi3VForCausalLM"),
}

Traceback (most recent call last):
File "/usr/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap
self.run()
File "/usr/lib/python3.10/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "/home/tarik/AI-WS/vllm/vllm/engine/multiprocessing/engine.py", line 397, in run_mp_engine
engine = MQLLMEngine.from_engine_args(engine_args=engine_args,
File "/home/tarik/AI-WS/vllm/vllm/engine/multiprocessing/engine.py", line 140, in from_engine_args
engine_config = engine_args.create_engine_config()
File "/home/tarik/AI-WS/vllm/vllm/engine/arg_utils.py", line 930, in create_engine_config
model_config = self.create_model_config()
File "/home/tarik/AI-WS/vllm/vllm/engine/arg_utils.py", line 864, in create_model_config
return ModelConfig(
File "/home/tarik/AI-WS/vllm/vllm/config.py", line 224, in init
supported_tasks, task = self._resolve_task(task, self.hf_config)
File "/home/tarik/AI-WS/vllm/vllm/config.py", line 307, in _resolve_task
raise ValueError(msg)
ValueError: This model does not support the 'embedding' task. Supported tasks: {'generate'}

please add embedding' task support
thank you

Alternatives

No response

Additional context

No response

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.

[DarkLight1337]

Qwen/Qwen2.5-1.5B-Instruct

This model is not designed for embeddings. What are you trying to do with this model?