registry.py 23.8 KB
Newer Older
1
# SPDX-License-Identifier: Apache-2.0
2
3
4
5
"""
Whenever you add an architecture to this page, please also update
`tests/models/registry.py` with example HuggingFace models for it.
"""
6
import importlib
7
import os
8
import pickle
9
10
import subprocess
import sys
11
import tempfile
12
13
14
from abc import ABC, abstractmethod
from dataclasses import dataclass, field
from functools import lru_cache
15
16
from typing import (AbstractSet, Callable, Dict, List, Optional, Tuple, Type,
                    TypeVar, Union)
17

18
import cloudpickle
19
20
21
import torch.nn as nn

from vllm.logger import init_logger
22
from vllm.utils import is_in_doc_build
23

24
25
26
27
from .interfaces import (has_inner_state, has_noops, is_attention_free,
                         is_hybrid, supports_cross_encoding,
                         supports_multimodal, supports_pp,
                         supports_transcription, supports_v0_only)
28
from .interfaces_base import is_text_generation_model
29
30
31

logger = init_logger(__name__)

32
# yapf: disable
33
34
_TEXT_GENERATION_MODELS = {
    # [Decoder-only]
35
36
37
    "AquilaModel": ("llama", "LlamaForCausalLM"),
    "AquilaForCausalLM": ("llama", "LlamaForCausalLM"),  # AquilaChat2
    "ArcticForCausalLM": ("arctic", "ArcticForCausalLM"),
38
    "MiniMaxText01ForCausalLM": ("minimax_text_01", "MiniMaxText01ForCausalLM"),
39
40
41
42
    # baichuan-7b, upper case 'C' in the class name
    "BaiChuanForCausalLM": ("baichuan", "BaiChuanForCausalLM"),
    # baichuan-13b, lower case 'c' in the class name
    "BaichuanForCausalLM": ("baichuan", "BaichuanForCausalLM"),
Yu Chin Fabian Lim's avatar
Yu Chin Fabian Lim committed
43
    "BambaForCausalLM": ("bamba", "BambaForCausalLM"),
44
    "BloomForCausalLM": ("bloom", "BloomForCausalLM"),
45
    "ChatGLMModel": ("chatglm", "ChatGLMForCausalLM"),
46
    "ChatGLMForConditionalGeneration": ("chatglm", "ChatGLMForCausalLM"),
47
    "CohereForCausalLM": ("commandr", "CohereForCausalLM"),
48
    "Cohere2ForCausalLM": ("commandr", "CohereForCausalLM"),
49
    "DbrxForCausalLM": ("dbrx", "DbrxForCausalLM"),
50
    "DeciLMForCausalLM": ("nemotron_nas", "DeciLMForCausalLM"),
51
52
    "DeepseekForCausalLM": ("deepseek", "DeepseekForCausalLM"),
    "DeepseekV2ForCausalLM": ("deepseek_v2", "DeepseekV2ForCausalLM"),
53
    "DeepseekV3ForCausalLM": ("deepseek_v2", "DeepseekV3ForCausalLM"),
54
55
    "ExaoneForCausalLM": ("exaone", "ExaoneForCausalLM"),
    "FalconForCausalLM": ("falcon", "FalconForCausalLM"),
56
    "Fairseq2LlamaForCausalLM": ("fairseq2_llama", "Fairseq2LlamaForCausalLM"),
57
58
    "GemmaForCausalLM": ("gemma", "GemmaForCausalLM"),
    "Gemma2ForCausalLM": ("gemma2", "Gemma2ForCausalLM"),
59
    "Gemma3ForCausalLM": ("gemma3", "Gemma3ForCausalLM"),
60
    "GlmForCausalLM": ("glm", "GlmForCausalLM"),
Yuxuan Zhang's avatar
Yuxuan Zhang committed
61
    "Glm4ForCausalLM": ("glm4", "Glm4ForCausalLM"),
62
63
64
65
66
67
    "GPT2LMHeadModel": ("gpt2", "GPT2LMHeadModel"),
    "GPTBigCodeForCausalLM": ("gpt_bigcode", "GPTBigCodeForCausalLM"),
    "GPTJForCausalLM": ("gpt_j", "GPTJForCausalLM"),
    "GPTNeoXForCausalLM": ("gpt_neox", "GPTNeoXForCausalLM"),
    "GraniteForCausalLM": ("granite", "GraniteForCausalLM"),
    "GraniteMoeForCausalLM": ("granitemoe", "GraniteMoeForCausalLM"),
68
    "GraniteMoeSharedForCausalLM": ("granitemoeshared", "GraniteMoeSharedForCausalLM"),   # noqa: E501
69
    "GritLM": ("gritlm", "GritLM"),
Michael Goin's avatar
Michael Goin committed
70
    "Grok1ModelForCausalLM": ("grok1", "Grok1ForCausalLM"),
71
72
    "InternLMForCausalLM": ("llama", "LlamaForCausalLM"),
    "InternLM2ForCausalLM": ("internlm2", "InternLM2ForCausalLM"),
73
    "InternLM2VEForCausalLM": ("internlm2_ve", "InternLM2VEForCausalLM"),
74
    "InternLM3ForCausalLM": ("llama", "LlamaForCausalLM"),
75
76
77
78
79
    "JAISLMHeadModel": ("jais", "JAISLMHeadModel"),
    "JambaForCausalLM": ("jamba", "JambaForCausalLM"),
    "LlamaForCausalLM": ("llama", "LlamaForCausalLM"),
    # For decapoda-research/llama-*
    "LLaMAForCausalLM": ("llama", "LlamaForCausalLM"),
80
    "MambaForCausalLM": ("mamba", "MambaForCausalLM"),
81
    "FalconMambaForCausalLM": ("mamba", "MambaForCausalLM"),
82
    "Mamba2ForCausalLM": ("mamba2", "Mamba2ForCausalLM"),
83
84
    "MiniCPMForCausalLM": ("minicpm", "MiniCPMForCausalLM"),
    "MiniCPM3ForCausalLM": ("minicpm3", "MiniCPM3ForCausalLM"),
85
86
87
88
89
90
91
92
    "MistralForCausalLM": ("llama", "LlamaForCausalLM"),
    "MixtralForCausalLM": ("mixtral", "MixtralForCausalLM"),
    "QuantMixtralForCausalLM": ("mixtral_quant", "MixtralForCausalLM"),
    # transformers's mpt class has lower case
    "MptForCausalLM": ("mpt", "MPTForCausalLM"),
    "MPTForCausalLM": ("mpt", "MPTForCausalLM"),
    "NemotronForCausalLM": ("nemotron", "NemotronForCausalLM"),
    "OlmoForCausalLM": ("olmo", "OlmoForCausalLM"),
93
    "Olmo2ForCausalLM": ("olmo2", "Olmo2ForCausalLM"),
94
95
96
97
98
99
100
101
    "OlmoeForCausalLM": ("olmoe", "OlmoeForCausalLM"),
    "OPTForCausalLM": ("opt", "OPTForCausalLM"),
    "OrionForCausalLM": ("orion", "OrionForCausalLM"),
    "PersimmonForCausalLM": ("persimmon", "PersimmonForCausalLM"),
    "PhiForCausalLM": ("phi", "PhiForCausalLM"),
    "Phi3ForCausalLM": ("phi3", "Phi3ForCausalLM"),
    "Phi3SmallForCausalLM": ("phi3_small", "Phi3SmallForCausalLM"),
    "PhiMoEForCausalLM": ("phimoe", "PhiMoEForCausalLM"),
Shinichi Hemmi's avatar
Shinichi Hemmi committed
102
    "Plamo2ForCausalLM": ("plamo2", "Plamo2ForCausalLM"),
103
    "QWenLMHeadModel": ("qwen", "QWenLMHeadModel"),
104
105
    "Qwen2ForCausalLM": ("qwen2", "Qwen2ForCausalLM"),
    "Qwen2MoeForCausalLM": ("qwen2_moe", "Qwen2MoeForCausalLM"),
106
107
    "Qwen3ForCausalLM": ("qwen3", "Qwen3ForCausalLM"),
    "Qwen3MoeForCausalLM": ("qwen3_moe", "Qwen3MoeForCausalLM"),
108
109
110
111
112
    "RWForCausalLM": ("falcon", "FalconForCausalLM"),
    "StableLMEpochForCausalLM": ("stablelm", "StablelmForCausalLM"),
    "StableLmForCausalLM": ("stablelm", "StablelmForCausalLM"),
    "Starcoder2ForCausalLM": ("starcoder2", "Starcoder2ForCausalLM"),
    "SolarForCausalLM": ("solar", "SolarForCausalLM"),
113
    "TeleChat2ForCausalLM": ("telechat2", "TeleChat2ForCausalLM"),
114
    "TeleFLMForCausalLM": ("teleflm", "TeleFLMForCausalLM"),
115
    "XverseForCausalLM": ("llama", "LlamaForCausalLM"),
116
    "Zamba2ForCausalLM": ("zamba2", "Zamba2ForCausalLM"),
117
118
119
    # [Encoder-decoder]
    "BartModel": ("bart", "BartForConditionalGeneration"),
    "BartForConditionalGeneration": ("bart", "BartForConditionalGeneration"),
120
121
122
}

_EMBEDDING_MODELS = {
123
    # [Text-only]
124
    "BertModel": ("bert", "BertEmbeddingModel"),
125
    "RobertaModel": ("roberta", "RobertaEmbeddingModel"),
126
    "RobertaForMaskedLM": ("roberta", "RobertaEmbeddingModel"),
127
    "XLMRobertaModel": ("roberta", "RobertaEmbeddingModel"),
128
    "DeciLMForCausalLM": ("nemotron_nas", "DeciLMForCausalLM"),
129
    "Gemma2Model": ("gemma2", "Gemma2ForCausalLM"),
130
    "GlmForCausalLM": ("glm", "GlmForCausalLM"),
131
    "GritLM": ("gritlm", "GritLM"),
132
    "InternLM2ForRewardModel": ("internlm2", "InternLM2ForRewardModel"),
133
    "JambaForSequenceClassification": ("jamba", "JambaForSequenceClassification"),  # noqa: E501
134
    "LlamaModel": ("llama", "LlamaForCausalLM"),
135
136
137
138
139
    **{
        # Multiple models share the same architecture, so we include them all
        k: (mod, arch) for k, (mod, arch) in _TEXT_GENERATION_MODELS.items()
        if arch == "LlamaForCausalLM"
    },
140
    "MistralModel": ("llama", "LlamaForCausalLM"),
141
    "Phi3ForCausalLM": ("phi3", "Phi3ForCausalLM"),
142
143
    "Qwen2Model": ("qwen2", "Qwen2EmbeddingModel"),
    "Qwen2ForCausalLM": ("qwen2", "Qwen2ForCausalLM"),
144
    "Qwen2ForRewardModel": ("qwen2_rm", "Qwen2ForRewardModel"),
145
    "Qwen2ForProcessRewardModel": ("qwen2_rm", "Qwen2ForProcessRewardModel"),
146
    "TeleChat2ForCausalLM": ("telechat2", "TeleChat2ForCausalLM"),
147
    # [Multimodal]
Cyrus Leung's avatar
Cyrus Leung committed
148
    "LlavaNextForConditionalGeneration": ("llava_next", "LlavaNextForConditionalGeneration"),  # noqa: E501
149
    "Phi3VForCausalLM": ("phi3v", "Phi3VForCausalLM"),
150
    "Qwen2VLForConditionalGeneration": ("qwen2_vl", "Qwen2VLForConditionalGeneration"),  # noqa: E501
151
152
    # [Auto-converted (see adapters.py)]
    "Qwen2ForSequenceClassification": ("qwen2", "Qwen2ForCausalLM"),
153
154
155
156
    # Technically PrithviGeoSpatialMAE is a model that works on images, both in
    # input and output. I am adding it here because it piggy-backs on embedding
    # models for the time being.
    "PrithviGeoSpatialMAE": ("prithvi_geospatial_mae", "PrithviGeoSpatialMAE"),
157
158
}

159
160
161
162
163
164
165
166
_CROSS_ENCODER_MODELS = {
    "BertForSequenceClassification": ("bert", "BertForSequenceClassification"),
    "RobertaForSequenceClassification": ("roberta",
                                         "RobertaForSequenceClassification"),
    "XLMRobertaForSequenceClassification": ("roberta",
                                            "RobertaForSequenceClassification"),
}

167
_MULTIMODAL_MODELS = {
168
    # [Decoder-only]
169
    "AriaForConditionalGeneration": ("aria", "AriaForConditionalGeneration"),
Jennifer Zhao's avatar
Jennifer Zhao committed
170
    "AyaVisionForConditionalGeneration": ("aya_vision", "AyaVisionForConditionalGeneration"),  # noqa: E501
171
172
    "Blip2ForConditionalGeneration": ("blip2", "Blip2ForConditionalGeneration"),
    "ChameleonForConditionalGeneration": ("chameleon", "ChameleonForConditionalGeneration"),  # noqa: E501
173
    "DeepseekVLV2ForCausalLM": ("deepseek_vl2", "DeepseekVLV2ForCausalLM"),
174
    "FuyuForCausalLM": ("fuyu", "FuyuForCausalLM"),
175
    "Gemma3ForConditionalGeneration": ("gemma3_mm", "Gemma3ForConditionalGeneration"),  # noqa: E501
176
    "GLM4VForCausalLM": ("glm4v", "GLM4VForCausalLM"),
177
    "H2OVLChatModel": ("h2ovl", "H2OVLChatModel"),
178
    "InternVLChatModel": ("internvl", "InternVLChatModel"),
179
    "Idefics3ForConditionalGeneration":("idefics3","Idefics3ForConditionalGeneration"),
180
    "SmolVLMForConditionalGeneration": ("smolvlm","SmolVLMForConditionalGeneration"),  # noqa: E501
181
    "KimiVLForConditionalGeneration": ("kimi_vl", "KimiVLForConditionalGeneration"),  # noqa: E501
182
183
184
185
    "LlavaForConditionalGeneration": ("llava", "LlavaForConditionalGeneration"),
    "LlavaNextForConditionalGeneration": ("llava_next", "LlavaNextForConditionalGeneration"),  # noqa: E501
    "LlavaNextVideoForConditionalGeneration": ("llava_next_video", "LlavaNextVideoForConditionalGeneration"),  # noqa: E501
    "LlavaOnevisionForConditionalGeneration": ("llava_onevision", "LlavaOnevisionForConditionalGeneration"),  # noqa: E501
186
    "MantisForConditionalGeneration": ("llava", "MantisForConditionalGeneration"),  # noqa: E501
187
    "MiniCPMO": ("minicpmo", "MiniCPMO"),
188
    "MiniCPMV": ("minicpmv", "MiniCPMV"),
189
    "Mistral3ForConditionalGeneration": ("mistral3", "Mistral3ForConditionalGeneration"),  # noqa: E501
190
    "MolmoForCausalLM": ("molmo", "MolmoForCausalLM"),
191
192
    "NVLM_D": ("nvlm_d", "NVLM_D_Model"),
    "PaliGemmaForConditionalGeneration": ("paligemma", "PaliGemmaForConditionalGeneration"),  # noqa: E501
193
    "Phi3VForCausalLM": ("phi3v", "Phi3VForCausalLM"),
194
    "PixtralForConditionalGeneration": ("pixtral", "PixtralForConditionalGeneration"),  # noqa: E501
195
    "QwenVLForConditionalGeneration": ("qwen_vl", "QwenVLForConditionalGeneration"),  # noqa: E501
196
    "Qwen2VLForConditionalGeneration": ("qwen2_vl", "Qwen2VLForConditionalGeneration"),  # noqa: E501
Roger Wang's avatar
Roger Wang committed
197
    "Qwen2_5_VLForConditionalGeneration": ("qwen2_5_vl", "Qwen2_5_VLForConditionalGeneration"),  # noqa: E501
198
    "Qwen2AudioForConditionalGeneration": ("qwen2_audio", "Qwen2AudioForConditionalGeneration"),  # noqa: E501
199
    "UltravoxModel": ("ultravox", "UltravoxModel"),
200
    "Phi4MMForCausalLM": ("phi4mm", "Phi4MMForCausalLM"),
201
    # [Encoder-decoder]
202
    "Florence2ForConditionalGeneration": ("florence2", "Florence2ForConditionalGeneration"),  # noqa: E501
203
    "MllamaForConditionalGeneration": ("mllama", "MllamaForConditionalGeneration"),  # noqa: E501
204
    "Llama4ForConditionalGeneration": ("mllama4", "Llama4ForConditionalGeneration"),  # noqa: E501
205
    "SkyworkR1VChatModel": ("skyworkr1v", "SkyworkR1VChatModel"),
206
    "WhisperForConditionalGeneration": ("whisper", "WhisperForConditionalGeneration"),  # noqa: E501
207
}
208
209
210

_SPECULATIVE_DECODING_MODELS = {
    "EAGLEModel": ("eagle", "EAGLE"),
211
    "EagleLlamaForCausalLM": ("llama_eagle", "EagleLlamaForCausalLM"),
212
    "DeepSeekMTPModel": ("deepseek_mtp", "DeepSeekMTP"),
213
214
    "MedusaModel": ("medusa", "Medusa"),
    "MLPSpeculatorPreTrainedModel": ("mlp_speculator", "MLPSpeculator"),
215
}
216

217
_TRANSFORMERS_MODELS = {
218
    "TransformersForCausalLM": ("transformers", "TransformersForCausalLM"),
219
}
220
# yapf: enable
221

222
_VLLM_MODELS = {
223
    **_TEXT_GENERATION_MODELS,
224
    **_EMBEDDING_MODELS,
225
    **_CROSS_ENCODER_MODELS,
226
    **_MULTIMODAL_MODELS,
227
    **_SPECULATIVE_DECODING_MODELS,
228
    **_TRANSFORMERS_MODELS,
229
230
}

231
232
233
234
235
236
237
238
# This variable is used as the args for subprocess.run(). We
# can modify  this variable to alter the args if needed. e.g.
# when we use par format to pack things together, sys.executable
# might not be the target we want to run.
_SUBPROCESS_COMMAND = [
    sys.executable, "-m", "vllm.model_executor.models.registry"
]

239

240
241
@dataclass(frozen=True)
class _ModelInfo:
242
    architecture: str
243
    is_text_generation_model: bool
244
    is_pooling_model: bool
245
    supports_cross_encoding: bool
246
247
    supports_multimodal: bool
    supports_pp: bool
248
249
    has_inner_state: bool
    is_attention_free: bool
250
    is_hybrid: bool
251
    has_noops: bool
252
    supports_transcription: bool
253
    supports_v0_only: bool
254
255

    @staticmethod
256
257
    def from_model_cls(model: Type[nn.Module]) -> "_ModelInfo":
        return _ModelInfo(
258
            architecture=model.__name__,
259
            is_text_generation_model=is_text_generation_model(model),
260
            is_pooling_model=True,  # Can convert any model into a pooling model
261
            supports_cross_encoding=supports_cross_encoding(model),
262
263
            supports_multimodal=supports_multimodal(model),
            supports_pp=supports_pp(model),
264
265
            has_inner_state=has_inner_state(model),
            is_attention_free=is_attention_free(model),
266
            is_hybrid=is_hybrid(model),
267
268
            supports_transcription=supports_transcription(model),
            supports_v0_only=supports_v0_only(model),
269
            has_noops=has_noops(model),
270
        )
271
272


273
class _BaseRegisteredModel(ABC):
274

275
276
277
    @abstractmethod
    def inspect_model_cls(self) -> _ModelInfo:
        raise NotImplementedError
278

279
280
281
    @abstractmethod
    def load_model_cls(self) -> Type[nn.Module]:
        raise NotImplementedError
282
283


284
285
286
287
288
289
290
291
@dataclass(frozen=True)
class _RegisteredModel(_BaseRegisteredModel):
    """
    Represents a model that has already been imported in the main process.
    """

    interfaces: _ModelInfo
    model_cls: Type[nn.Module]
292
293

    @staticmethod
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
    def from_model_cls(model_cls: Type[nn.Module]):
        return _RegisteredModel(
            interfaces=_ModelInfo.from_model_cls(model_cls),
            model_cls=model_cls,
        )

    def inspect_model_cls(self) -> _ModelInfo:
        return self.interfaces

    def load_model_cls(self) -> Type[nn.Module]:
        return self.model_cls


@dataclass(frozen=True)
class _LazyRegisteredModel(_BaseRegisteredModel):
    """
    Represents a model that has not been imported in the main process.
    """
    module_name: str
    class_name: str

    # Performed in another process to avoid initializing CUDA
    def inspect_model_cls(self) -> _ModelInfo:
        return _run_in_subprocess(
            lambda: _ModelInfo.from_model_cls(self.load_model_cls()))

    def load_model_cls(self) -> Type[nn.Module]:
        mod = importlib.import_module(self.module_name)
        return getattr(mod, self.class_name)


@lru_cache(maxsize=128)
def _try_load_model_cls(
    model_arch: str,
    model: _BaseRegisteredModel,
) -> Optional[Type[nn.Module]]:
330
    from vllm.platforms import current_platform
331
    current_platform.verify_model_arch(model_arch)
332
333
334
335
336
337
    try:
        return model.load_model_cls()
    except Exception:
        logger.exception("Error in loading model architecture '%s'",
                         model_arch)
        return None
338
339


340
341
342
343
344
345
346
347
348
349
350
@lru_cache(maxsize=128)
def _try_inspect_model_cls(
    model_arch: str,
    model: _BaseRegisteredModel,
) -> Optional[_ModelInfo]:
    try:
        return model.inspect_model_cls()
    except Exception:
        logger.exception("Error in inspecting model architecture '%s'",
                         model_arch)
        return None
351
352


353
354
355
356
@dataclass
class _ModelRegistry:
    # Keyed by model_arch
    models: Dict[str, _BaseRegisteredModel] = field(default_factory=dict)
357

358
359
    def get_supported_archs(self) -> AbstractSet[str]:
        return self.models.keys()
360

361
362
363
364
365
    def register_model(
        self,
        model_arch: str,
        model_cls: Union[Type[nn.Module], str],
    ) -> None:
366
367
368
369
370
371
372
373
374
375
376
        """
        Register an external model to be used in vLLM.

        :code:`model_cls` can be either:

        - A :class:`torch.nn.Module` class directly referencing the model.
        - A string in the format :code:`<module>:<class>` which can be used to
          lazily import the model. This is useful to avoid initializing CUDA
          when importing the model and thus the related error
          :code:`RuntimeError: Cannot re-initialize CUDA in forked subprocess`.
        """
377
378
379
380
        if not isinstance(model_arch, str):
            msg = f"`model_arch` should be a string, not a {type(model_arch)}"
            raise TypeError(msg)

381
        if model_arch in self.models:
382
383
384
            logger.warning(
                "Model architecture %s is already registered, and will be "
                "overwritten by the new model class %s.", model_arch,
385
386
387
388
389
390
391
                model_cls)

        if isinstance(model_cls, str):
            split_str = model_cls.split(":")
            if len(split_str) != 2:
                msg = "Expected a string in the format `<module>:<class>`"
                raise ValueError(msg)
392

393
            model = _LazyRegisteredModel(*split_str)
394
395
        elif isinstance(model_cls, type) and (is_in_doc_build() or issubclass(
                model_cls, nn.Module)):
396
            model = _RegisteredModel.from_model_cls(model_cls)
397
398
399
400
        else:
            msg = ("`model_cls` should be a string or PyTorch model class, "
                   f"not a {type(model_arch)}")
            raise TypeError(msg)
401

402
        self.models[model_arch] = model
403

404
405
    def _raise_for_unsupported(self, architectures: List[str]):
        all_supported_archs = self.get_supported_archs()
406

407
408
409
410
411
        if any(arch in all_supported_archs for arch in architectures):
            raise ValueError(
                f"Model architectures {architectures} failed "
                "to be inspected. Please check the logs for more details.")

412
413
414
        raise ValueError(
            f"Model architectures {architectures} are not supported for now. "
            f"Supported architectures: {all_supported_archs}")
415

416
417
418
419
    def _try_load_model_cls(self,
                            model_arch: str) -> Optional[Type[nn.Module]]:
        if model_arch not in self.models:
            return None
420

421
        return _try_load_model_cls(model_arch, self.models[model_arch])
422

423
424
425
    def _try_inspect_model_cls(self, model_arch: str) -> Optional[_ModelInfo]:
        if model_arch not in self.models:
            return None
426

427
        return _try_inspect_model_cls(model_arch, self.models[model_arch])
428

429
430
431
432
    def _normalize_archs(
        self,
        architectures: Union[str, List[str]],
    ) -> List[str]:
433
434
435
436
437
        if isinstance(architectures, str):
            architectures = [architectures]
        if not architectures:
            logger.warning("No model architectures are specified")

438
439
440
441
        # filter out support architectures
        normalized_arch = list(
            filter(lambda model: model in self.models, architectures))

442
        # make sure Transformers backend is put at the last as a fallback
443
        if len(normalized_arch) != len(architectures):
444
            normalized_arch.append("TransformersForCausalLM")
445
        return normalized_arch
446

447
448
449
    def inspect_model_cls(
        self,
        architectures: Union[str, List[str]],
450
    ) -> Tuple[_ModelInfo, str]:
451
        architectures = self._normalize_archs(architectures)
452

453
454
455
        for arch in architectures:
            model_info = self._try_inspect_model_cls(arch)
            if model_info is not None:
456
                return (model_info, arch)
457

458
        return self._raise_for_unsupported(architectures)
459

460
461
462
463
464
    def resolve_model_cls(
        self,
        architectures: Union[str, List[str]],
    ) -> Tuple[Type[nn.Module], str]:
        architectures = self._normalize_archs(architectures)
465

466
467
468
469
        for arch in architectures:
            model_cls = self._try_load_model_cls(arch)
            if model_cls is not None:
                return (model_cls, arch)
470

471
        return self._raise_for_unsupported(architectures)
472

473
474
475
476
    def is_text_generation_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
477
478
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.is_text_generation_model
479

480
    def is_pooling_model(
481
482
483
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
484
        model_cls, _ = self.inspect_model_cls(architectures)
485
        return model_cls.is_pooling_model
486

487
488
489
490
    def is_cross_encoder_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
491
492
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.supports_cross_encoding
493

494
495
496
497
    def is_multimodal_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
498
499
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.supports_multimodal
500
501
502
503
504

    def is_pp_supported_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
505
506
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.supports_pp
507

508
509
510
511
512
513
    def model_has_inner_state(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.has_inner_state
514

515
516
517
518
519
520
    def is_attention_free_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.is_attention_free
521

522
523
524
525
526
527
528
    def is_hybrid_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.is_hybrid

529
530
531
532
533
534
535
    def is_noops_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.has_noops

536
537
538
539
540
541
542
    def is_transcription_model(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
        model_cls, _ = self.inspect_model_cls(architectures)
        return model_cls.supports_transcription

543
544
545
546
547
548
549
    def is_v1_compatible(
        self,
        architectures: Union[str, List[str]],
    ) -> bool:
        model_cls, _ = self.inspect_model_cls(architectures)
        return not model_cls.supports_v0_only

550
551

ModelRegistry = _ModelRegistry({
552
553
    model_arch:
    _LazyRegisteredModel(
554
555
556
557
558
559
560
561
562
563
        module_name=f"vllm.model_executor.models.{mod_relname}",
        class_name=cls_name,
    )
    for model_arch, (mod_relname, cls_name) in _VLLM_MODELS.items()
})

_T = TypeVar("_T")


def _run_in_subprocess(fn: Callable[[], _T]) -> _T:
564
565
566
567
568
    # NOTE: We use a temporary directory instead of a temporary file to avoid
    # issues like https://stackoverflow.com/questions/23212435/permission-denied-to-write-to-my-temporary-file
    with tempfile.TemporaryDirectory() as tempdir:
        output_filepath = os.path.join(tempdir, "registry_output.tmp")

569
        # `cloudpickle` allows pickling lambda functions directly
570
        input_bytes = cloudpickle.dumps((fn, output_filepath))
571
572
573

        # cannot use `sys.executable __file__` here because the script
        # contains relative imports
574
575
576
        returned = subprocess.run(_SUBPROCESS_COMMAND,
                                  input=input_bytes,
                                  capture_output=True)
577
578
579
580
581
582
583
584
585

        # check if the subprocess is successful
        try:
            returned.check_returncode()
        except Exception as e:
            # wrap raised exception to provide more information
            raise RuntimeError(f"Error raised in subprocess:\n"
                               f"{returned.stderr.decode()}") from e

586
        with open(output_filepath, "rb") as f:
587
588
589
590
591
592
593
594
595
596
597
            return pickle.load(f)


def _run() -> None:
    # Setup plugins
    from vllm.plugins import load_general_plugins
    load_general_plugins()

    fn, output_file = pickle.loads(sys.stdin.buffer.read())

    result = fn()
598
599
600

    with open(output_file, "wb") as f:
        f.write(pickle.dumps(result))
601
602
603


if __name__ == "__main__":
604
    _run()