Qwen

/

Qwen3.5-27B

like 953

Follow

Qwen 80.5k

python -c "import transformers; print(transformers.version)"
5.3.0.dev0

print("Initializing vLLM Engine...")
engine_args = AsyncEngineArgs(
model=MODEL_PATH,
tokenizer=MODEL_PATH,
trust_remote_code=True,
tensor_parallel_size=1,
dtype="bfloat16",
quantization="bitsandbytes",
max_model_len=8192,
gpu_memory_utilization=0.85,
enforce_eager=False,
max_num_seqs=4,
)
vllm_engine = AsyncLLMEngine.from_engine_args(engine_args)

Initializing vLLM Engine...

ValidationError Traceback (most recent call last)
Cell In[4], line 14
1 print("Initializing vLLM Engine...")
2 engine_args = AsyncEngineArgs(
3 model=MODEL_PATH,
4 tokenizer=MODEL_PATH,
(...) 12 max_num_seqs=4,
13 )
---> 14 vllm_engine = AsyncLLMEngine.from_engine_args(engine_args)

File ~\miniconda3\envs\vllm_clean\Lib\site-packages\vllm\v1\engine\async_llm.py:231, in AsyncLLM.from_engine_args(cls, engine_args, start_engine_loop, usage_context, stat_loggers)
228 """Create an AsyncLLM from the EngineArgs."""
230 # Create the engine configs.
--> 231 vllm_config = engine_args.create_engine_config(usage_context)
232 executor_class = Executor.get_class(vllm_config)
234 # Create the AsyncLLM.

File ~\miniconda3\envs\vllm_clean\Lib\site-packages\vllm\engine\arg_utils.py:1142, in EngineArgs.create_engine_config(self, usage_context, headless)
1131 device_config = DeviceConfig(
1132 device=cast(Device, current_platform.device_type))
1134 (self.model, self.tokenizer,
1135 self.speculative_config) = maybe_override_with_speculators(
1136 model=self.model,
(...) 1140 vllm_speculative_config=self.speculative_config,
1141 )
-> 1142 model_config = self.create_model_config()
1144 # * If VLLM_USE_V1 is unset, we enable V1 for "supported features"
1145 # and fall back to V0 for experimental or unsupported features.
1146 # * If VLLM_USE_V1=1, we enable V1 for supported + experimental
1147 # features and raise error for unsupported features.
1148 # * If VLLM_USE_V1=0, we disable V1.
1149 use_v1 = False

File ~\miniconda3\envs\vllm_clean\Lib\site-packages\vllm\engine\arg_utils.py:994, in EngineArgs.create_model_config(self)
987 logger.warning(
988 "--enable-multimodal-encoder-data-parallelis deprecated " 989 "and will be removed in v0.13. " 990 "Please use--mm-encoder-tp-mode data` instead.")
992 self.mm_encoder_tp_mode = "data"
--> 994 return ModelConfig(
995 model=self.model,
996 hf_config_path=self.hf_config_path,
997 runner=self.runner,
998 convert=self.convert,
999 task=self.task,
1000 tokenizer=self.tokenizer,
1001 tokenizer_mode=self.tokenizer_mode,
1002 trust_remote_code=self.trust_remote_code,
1003 allowed_local_media_path=self.allowed_local_media_path,
1004 allowed_media_domains=self.allowed_media_domains,
1005 dtype=self.dtype,
1006 seed=self.seed,
1007 revision=self.revision,
1008 code_revision=self.code_revision,
1009 rope_scaling=self.rope_scaling,
1010 rope_theta=self.rope_theta,
1011 hf_token=self.hf_token,
1012 hf_overrides=self.hf_overrides,
1013 tokenizer_revision=self.tokenizer_revision,
1014 max_model_len=self.max_model_len,
1015 quantization=self.quantization,
1016 enforce_eager=self.enforce_eager,
1017 max_logprobs=self.max_logprobs,
1018 logprobs_mode=self.logprobs_mode,
1019 disable_sliding_window=self.disable_sliding_window,
1020 disable_cascade_attn=self.disable_cascade_attn,
1021 skip_tokenizer_init=self.skip_tokenizer_init,
1022 enable_prompt_embeds=self.enable_prompt_embeds,
1023 served_model_name=self.served_model_name,
1024 limit_mm_per_prompt=self.limit_mm_per_prompt,
1025 interleave_mm_strings=self.interleave_mm_strings,
1026 media_io_kwargs=self.media_io_kwargs,
1027 skip_mm_profiling=self.skip_mm_profiling,
1028 config_format=self.config_format,
1029 mm_processor_kwargs=self.mm_processor_kwargs,
1030 mm_processor_cache_gb=self.mm_processor_cache_gb,
1031 mm_processor_cache_type=self.mm_processor_cache_type,
1032 mm_shm_cache_max_object_size_mb=self.
1033 mm_shm_cache_max_object_size_mb,
1034 mm_encoder_tp_mode=self.mm_encoder_tp_mode,
1035 pooler_config=self.pooler_config,
1036 override_pooler_config=self.override_pooler_config,
1037 logits_processor_pattern=self.logits_processor_pattern,
1038 generation_config=self.generation_config,
1039 override_generation_config=self.override_generation_config,
1040 enable_sleep_mode=self.enable_sleep_mode,
1041 model_impl=self.model_impl,
1042 override_attention_dtype=self.override_attention_dtype,
1043 logits_processors=self.logits_processors,
1044 video_pruning_rate=self.video_pruning_rate,
1045 io_processor_plugin=self.io_processor_plugin,
1046 )

File ~\miniconda3\envs\vllm_clean\Lib\site-packages\pydantic_internal_dataclasses.py:121, in complete_dataclass..init(dataclass_self, *args, **kwargs)
119 tracebackhide = True
120 s = dataclass_self
--> 121 s.pydantic_validator.validate_python(ArgsKwargs(args, kwargs), self_instance=s)

ValidationError: 1 validation error for ModelConfig
Value error, Model architectures ['Qwen3_5ForConditionalGeneration'] are not supported for now. Supported architectures: dict_keys(['ApertusForCausalLM', 'AquilaModel', 'AquilaForCausalLM', 'ArceeForCausalLM', 'ArcticForCausalLM', 'MiniMaxForCausalLM', 'MiniMaxText01ForCausalLM', 'MiniMaxM1ForCausalLM', 'BaiChuanForCausalLM', 'BaichuanForCausalLM', 'BailingMoeForCausalLM', 'BailingMoeV2ForCausalLM', 'BambaForCausalLM', 'BloomForCausalLM', 'ChatGLMModel', 'ChatGLMForConditionalGeneration', 'CohereForCausalLM', 'Cohere2ForCausalLM', 'CwmForCausalLM', 'DbrxForCausalLM', 'DeciLMForCausalLM', 'DeepseekForCausalLM', 'DeepseekV2ForCausalLM', 'DeepseekV3ForCausalLM', 'DeepseekV32ForCausalLM', 'Dots1ForCausalLM', 'Ernie4_5ForCausalLM', 'Ernie4_5_MoeForCausalLM', 'ExaoneForCausalLM', 'Exaone4ForCausalLM', 'FalconForCausalLM', 'Fairseq2LlamaForCausalLM', 'GemmaForCausalLM', 'Gemma2ForCausalLM', 'Gemma3ForCausalLM', 'Gemma3nForCausalLM', 'Qwen3NextForCausalLM', 'GlmForCausalLM', 'Glm4ForCausalLM', 'Glm4MoeForCausalLM', 'GptOssForCausalLM', 'GPT2LMHeadModel', 'GPTBigCodeForCausalLM', 'GPTJForCausalLM', 'GPTNeoXForCausalLM', 'GraniteForCausalLM', 'GraniteMoeForCausalLM', 'GraniteMoeHybridForCausalLM', 'GraniteMoeSharedForCausalLM', 'GritLM', 'Grok1ModelForCausalLM', 'HunYuanMoEV1ForCausalLM', 'HunYuanDenseV1ForCausalLM', 'HCXVisionForCausalLM', 'InternLMForCausalLM', 'InternLM2ForCausalLM', 'InternLM2VEForCausalLM', 'InternLM3ForCausalLM', 'JAISLMHeadModel', 'JambaForCausalLM', 'Lfm2ForCausalLM', 'LlamaForCausalLM', 'Llama4ForCausalLM', 'LLaMAForCausalLM', 'LongcatFlashForCausalLM', 'MambaForCausalLM', 'FalconMambaForCausalLM', 'FalconH1ForCausalLM', 'Mamba2ForCausalLM', 'MiniCPMForCausalLM', 'MiniCPM3ForCausalLM', 'MistralForCausalLM', 'MixtralForCausalLM', 'MotifForCausalLM', 'MptForCausalLM', 'MPTForCausalLM', 'MiMoForCausalLM', 'NemotronForCausalLM', 'NemotronHForCausalLM', 'OlmoForCausalLM', 'Olmo2ForCausalLM', 'Olmo3ForCausalLM', 'OlmoeForCausalLM', 'OPTForCausalLM', 'OrionForCausalLM', 'PersimmonForCausalLM', 'PhiForCausalLM', 'Phi3ForCausalLM', 'PhiMoEForCausalLM', 'Plamo2ForCausalLM', 'QWenLMHeadModel', 'Qwen2ForCausalLM', 'Qwen2MoeForCausalLM', 'Qwen3ForCausalLM', 'Qwen3MoeForCausalLM', 'RWForCausalLM', 'SeedOssForCausalLM', 'Step3TextForCausalLM', 'StableLMEpochForCausalLM', 'StableLmForCausalLM', 'Starcoder2ForCausalLM', 'SolarForCausalLM', 'TeleChat2ForCausalLM', 'TeleFLMForCausalLM', 'XverseForCausalLM', 'Zamba2ForCausalLM', 'BertModel', 'Gemma2Model', 'Gemma3TextModel', 'GPT2ForSequenceClassification', 'GteModel', 'GteNewModel', 'InternLM2ForRewardModel', 'JambaForSequenceClassification', 'LlamaModel', 'MistralModel', 'ModernBertModel', 'NomicBertModel', 'Qwen2Model', 'Qwen2ForRewardModel', 'Qwen2ForProcessRewardModel', 'RobertaForMaskedLM', 'RobertaModel', 'XLMRobertaModel', 'LlavaNextForConditionalGeneration', 'Phi3VForCausalLM', 'Qwen2VLForConditionalGeneration', 'PrithviGeoSpatialMAE', 'Terratorch', 'BertForSequenceClassification', 'BertForTokenClassification', 'GteNewForSequenceClassification', 'ModernBertForSequenceClassification', 'RobertaForSequenceClassification', 'XLMRobertaForSequenceClassification', 'JinaVLForRanking', 'AriaForConditionalGeneration', 'AyaVisionForConditionalGeneration', 'Blip2ForConditionalGeneration', 'ChameleonForConditionalGeneration', 'Cohere2VisionForConditionalGeneration', 'DeepseekVLV2ForCausalLM', 'DotsOCRForCausalLM', 'Ernie4_5_VLMoeForConditionalGeneration', 'FuyuForCausalLM', 'Gemma3ForConditionalGeneration', 'Gemma3nForConditionalGeneration', 'GLM4VForCausalLM', 'Glm4vForConditionalGeneration', 'Glm4vMoeForConditionalGeneration', 'GraniteSpeechForConditionalGeneration', 'H2OVLChatModel', 'InternVLChatModel', 'NemotronH_Nano_VL_V2', 'InternS1ForConditionalGeneration', 'InternVLForConditionalGeneration', 'Idefics3ForConditionalGeneration', 'SmolVLMForConditionalGeneration', 'KeyeForConditionalGeneration', 'KeyeVL1_5ForConditionalGeneration', 'RForConditionalGeneration', 'KimiVLForConditionalGeneration', 'Llama_Nemotron_Nano_VL', 'Llama4ForConditionalGeneration', 'LlavaForConditionalGeneration', 'LlavaNextVideoForConditionalGeneration', 'LlavaOnevisionForConditionalGeneration', 'MantisForConditionalGeneration', 'MiDashengLMModel', 'MiniMaxVL01ForConditionalGeneration', 'MiniCPMO', 'MiniCPMV', 'Mistral3ForConditionalGeneration', 'MolmoForCausalLM', 'NVLM_D', 'Ovis', 'Ovis2_5', 'PaliGemmaForConditionalGeneration', 'Phi4MMForCausalLM', 'Phi4MultimodalForCausalLM', 'PixtralForConditionalGeneration', 'QwenVLForConditionalGeneration', 'Qwen2_5_VLForConditionalGeneration', 'Qwen2AudioForConditionalGeneration', 'Qwen2_5OmniModel', 'Qwen2_5OmniForConditionalGeneration', 'Qwen3VLForConditionalGeneration', 'Qwen3VLMoeForConditionalGeneration', 'SkyworkR1VChatModel', 'Step3VLForConditionalGeneration', 'TarsierForConditionalGeneration', 'Tarsier2ForConditionalGeneration', 'UltravoxModel', 'VoxtralForConditionalGeneration', 'WhisperForConditionalGeneration', 'MiMoMTPModel', 'EagleLlamaForCausalLM', 'EagleLlama4ForCausalLM', 'EagleMiniCPMForCausalLM', 'Eagle3LlamaForCausalLM', 'LlamaForCausalLMEagle3', 'EagleDeepSeekMTPModel', 'DeepSeekMTPModel', 'ErnieMTPModel', 'LongCatFlashMTPModel', 'Glm4MoeMTPModel', 'MedusaModel', 'Qwen3NextMTP', 'SmolLM3ForCausalLM', 'Emu3ForConditionalGeneration', 'TransformersModel', 'TransformersForCausalLM', 'TransformersForMultimodalLM']) [type=value_error, input_value=ArgsKwargs((), {'model': ...rocessor_plugin': None}), input_type=ArgsKwargs]
For further information visit https://errors.pydantic.dev/2.12/v/value_error