Auto ν΄λμ€[[auto-classes]]
λ§μ κ²½μ°, μ¬μ©νλ €λ μν€ν
μ²λ from_pretrained() λ©μλμμ μ 곡νλ μ¬μ νλ ¨λ λͺ¨λΈμ μ΄λ¦μ΄λ κ²½λ‘λ‘λΆν° μ μΆν μ μμ΅λλ€. AutoClassesλ μ΄ μμ
μ μν΄ μ‘΄μ¬νλ©°, μ¬μ νμ΅λ λͺ¨λΈ κ°μ€μΉ/ꡬμ±/λ¨μ΄μ¬μ μ λν μ΄λ¦/κ²½λ‘λ₯Ό μ 곡νλ©΄ μλμΌλ‘ κ΄λ ¨ λͺ¨λΈμ κ°μ Έμ€λλ‘ λμμ€λλ€.
[AutoConfig], [AutoModel], [AutoTokenizer] μ€ νλλ₯Ό μΈμ€ν΄μ€ννλ©΄ ν΄λΉ μν€ν
μ²μ ν΄λμ€λ₯Ό μ§μ μμ±ν©λλ€. μλ₯Ό λ€μ΄,
model = AutoModel.from_pretrained("google-bert/bert-base-cased")
μ μ½λλ [BertModel]μ μΈμ€ν΄μ€μΈ λͺ¨λΈμ μμ±ν©λλ€.
κ° μμ
μ λν΄ νλμ AutoModel ν΄λμ€κ° μμΌλ©°, κ°κ°μ λ°±μλ(PyTorch, TensorFlow λλ Flax)μ ν΄λΉνλ ν΄λμ€κ° μ‘΄μ¬ν©λλ€.
μλ ν΄λμ€ νμ₯[[extending-the-auto-classes]]
κ° μλ ν΄λμ€λ μ¬μ©μμ 컀μ€ν
ν΄λμ€λ‘ νμ₯λ μ μλ λ©μλλ₯Ό κ°μ§κ³ μμ΅λλ€. μλ₯Ό λ€μ΄, NewModelμ΄λΌλ 컀μ€ν
λͺ¨λΈ ν΄λμ€λ₯Ό μ μνλ€λ©΄, NewModelConfigλ₯Ό μ€λΉν ν λ€μκ³Ό κ°μ΄ μλ ν΄λμ€μ μΆκ°ν μ μμ΅λλ€:
from transformers import AutoConfig, AutoModel
AutoConfig.register("new-model", NewModelConfig)
AutoModel.register(NewModelConfig, NewModel)
μ΄νμλ μΌλ°μ μΌλ‘ μλ ν΄λμ€λ₯Ό μ¬μ©νλ κ²μ²λΌ μ¬μ©ν μ μμ΅λλ€!
λ§μ½ NewModelConfigκ° [~transformers.PreTrainedConfig]μ μλΈν΄λμ€λΌλ©΄, ν΄λΉ model_type μμ±μ΄ λ±λ‘ν λ μ¬μ©νλ ν€(μ¬κΈ°μλ "new-model")μ λμΌνκ² μ€μ λμ΄ μλμ§ νμΈνμΈμ.
λ§μ°¬κ°μ§λ‘, NewModelμ΄ [PreTrainedModel]μ μλΈν΄λμ€λΌλ©΄, ν΄λΉ config_class μμ±μ΄ λ±λ‘ν λ μ¬μ©νλ ν΄λμ€(μ¬κΈ°μλ NewModelConfig)μ λμΌνκ² μ€μ λμ΄ μλμ§ νμΈνμΈμ.
AutoConfig[[transformers.AutoConfig]]
[[autodoc]] AutoConfig
AutoTokenizer[[transformers.AutoTokenizer]]
[[autodoc]] AutoTokenizer
AutoFeatureExtractor[[transformers.AutoFeatureExtractor]]
[[autodoc]] AutoFeatureExtractor
AutoImageProcessor[[transformers.AutoImageProcessor]]
[[autodoc]] AutoImageProcessor
AutoProcessor[[transformers.AutoProcessor]]
[[autodoc]] AutoProcessor
μΌλ°μ μΈ λͺ¨λΈ ν΄λμ€[[generic-model-classes]]
λ€μ μλ ν΄λμ€λ€μ νΉμ ν€λ μμ΄ κΈ°λ³Έ λͺ¨λΈ ν΄λμ€λ₯Ό μΈμ€ν΄μ€ννλ λ° μ¬μ©ν μ μμ΅λλ€.
AutoModel[[transformers.AutoModel]]
[[autodoc]] AutoModel
μΌλ°μ μΈ μ¬μ νμ΅ ν΄λμ€[[generic-pretraining-classes]]
λ€μ μλ ν΄λμ€λ€μ μ¬μ νλ ¨ ν€λκ° ν¬ν¨λ λͺ¨λΈμ μΈμ€ν΄μ€ννλ λ° μ¬μ©ν μ μμ΅λλ€.
AutoModelForPreTraining[[transformers.AutoModelForPreTraining]]
[[autodoc]] AutoModelForPreTraining
μμ°μ΄ μ²λ¦¬[[natural-language-processing]]
λ€μ μλ ν΄λμ€λ€μ μλμ μμ°μ΄ μ²λ¦¬ μμ μ μ¬μ©ν μ μμ΅λλ€.
AutoModelForCausalLM[[transformers.AutoModelForCausalLM]]
[[autodoc]] AutoModelForCausalLM
AutoModelForMaskedLM[[transformers.AutoModelForMaskedLM]]
[[autodoc]] AutoModelForMaskedLM
AutoModelForMaskGeneration[[transformers.AutoModelForMaskGeneration]]
[[autodoc]] AutoModelForMaskGeneration
AutoModelForSeq2SeqLM[[transformers.AutoModelForSeq2SeqLM]]
[[autodoc]] AutoModelForSeq2SeqLM
AutoModelForSequenceClassification[[transformers.AutoModelForSequenceClassification]]
[[autodoc]] AutoModelForSequenceClassification
AutoModelForMultipleChoice[[transformers.AutoModelForMultipleChoice]]
[[autodoc]] AutoModelForMultipleChoice
AutoModelForNextSentencePrediction[[transformers.AutoModelForNextSentencePrediction]]
[[autodoc]] AutoModelForNextSentencePrediction
AutoModelForTokenClassification[[transformers.AutoModelForTokenClassification]]
[[autodoc]] AutoModelForTokenClassification
AutoModelForQuestionAnswering[[transformers.AutoModelForQuestionAnswering]]
[[autodoc]] AutoModelForQuestionAnswering
AutoModelForTextEncoding[[transformers.AutoModelForTextEncoding]]
[[autodoc]] AutoModelForTextEncoding
μ»΄ν¨ν° λΉμ [[computer-vision]]
λ€μ μλ ν΄λμ€λ€μ μλμ μ»΄ν¨ν° λΉμ μμ μ μ¬μ©ν μ μμ΅λλ€.
AutoModelForDepthEstimation[[transformers.AutoModelForDepthEstimation]]
[[autodoc]] AutoModelForDepthEstimation
AutoModelForImageClassification[[transformers.AutoModelForImageClassification]]
[[autodoc]] AutoModelForImageClassification
AutoModelForVideoClassification[[transformers.AutoModelForVideoClassification]]
[[autodoc]] AutoModelForVideoClassification
AutoModelForKeypointDetection[[transformers.AutoModelForKeypointDetection]]
[[autodoc]] AutoModelForKeypointDetection
AutoModelForMaskedImageModeling[[transformers.AutoModelForMaskedImageModeling]]
[[autodoc]] AutoModelForMaskedImageModeling
AutoModelForObjectDetection[[transformers.AutoModelForObjectDetection]]
[[autodoc]] AutoModelForObjectDetection
AutoModelForImageSegmentation[[transformers.AutoModelForImageSegmentation]]
[[autodoc]] AutoModelForImageSegmentation
AutoModelForImageToImage[[transformers.AutoModelForImageToImage]]
[[autodoc]] AutoModelForImageToImage
AutoModelForSemanticSegmentation[[transformers.AutoModelForSemanticSegmentation]]
[[autodoc]] AutoModelForSemanticSegmentation
AutoModelForInstanceSegmentation[[transformers.AutoModelForInstanceSegmentation]]
[[autodoc]] AutoModelForInstanceSegmentation
AutoModelForUniversalSegmentation[[transformers.AutoModelForUniversalSegmentation]]
[[autodoc]] AutoModelForUniversalSegmentation
AutoModelForZeroShotImageClassification[[transformers.AutoModelForZeroShotImageClassification]]
[[autodoc]] AutoModelForZeroShotImageClassification
AutoModelForZeroShotObjectDetection[[transformers.AutoModelForZeroShotObjectDetection]]
[[autodoc]] AutoModelForZeroShotObjectDetection
μ€λμ€[[audio]]
λ€μ μλ ν΄λμ€λ€μ μλμ μ€λμ€ μμ μ μ¬μ©ν μ μμ΅λλ€.
AutoModelForAudioClassification[[transformers.AutoModelForAudioClassification]]
[[autodoc]] AutoModelForAudioClassification
AutoModelForAudioFrameClassification[[transformers.AutoModelForAudioFrameClassification]]
[[autodoc]] AutoModelForAudioFrameClassification
AutoModelForCTC[[transformers.AutoModelForCTC]]
[[autodoc]] AutoModelForCTC
AutoModelForSpeechSeq2Seq[[transformers.AutoModelForSpeechSeq2Seq]]
[[autodoc]] AutoModelForSpeechSeq2Seq
AutoModelForAudioXVector[[transformers.AutoModelForAudioXVector]]
[[autodoc]] AutoModelForAudioXVector
AutoModelForTextToSpectrogram[[transformers.AutoModelForTextToSpectrogram]]
[[autodoc]] AutoModelForTextToSpectrogram
AutoModelForTextToWaveform[[transformers.AutoModelForTextToWaveform]]
[[autodoc]] AutoModelForTextToWaveform
λ©ν°λͺ¨λ¬[[multimodal]]
λ€μ μλ ν΄λμ€λ€μ μλμ λ©ν°λͺ¨λ¬ μμ μ μ¬μ©ν μ μμ΅λλ€.
AutoModelForTableQuestionAnswering[[transformers.AutoModelForTableQuestionAnswering]]
[[autodoc]] AutoModelForTableQuestionAnswering
AutoModelForDocumentQuestionAnswering[[transformers.AutoModelForDocumentQuestionAnswering]]
[[autodoc]] AutoModelForDocumentQuestionAnswering
AutoModelForVisualQuestionAnswering[[transformers.AutoModelForVisualQuestionAnswering]]
[[autodoc]] AutoModelForVisualQuestionAnswering
Time Series
AutoModelForTimeSeriesPrediction[[transformers.AutoModelForTimeSeriesPrediction]]
[[autodoc]] AutoModelForTimeSeriesPrediction