Instancie uma das classes de configuração da biblioteca a partir da configuração de um modelo pré-treinado.

A classe de configuração a ser instanciada é selecionada com base na propriedade model_type do objeto de configuração que é carregado, ou, quando essa propriedade não existe, recorrendo ao uso de correspondência de padrões em pretrained_model_name_or_path:

• EvollaModel — EvollaConfig (modelo EvollaConfig)
• afmoe — AfmoeConfig (modelo AfmoeConfig)
• aimv2 — Aimv2Config (modelo Aimv2Config)
• aimv2_text_model — Aimv2TextConfig (modelo Aimv2TextConfig)
• aimv2_vision_model — Aimv2VisionConfig (modelo Aimv2VisionConfig)
• albert — AlbertConfig (modelo AlbertConfig)
• align — AlignConfig (modelo AlignConfig)
• align_text_model — AlignTextConfig (modelo AlignTextConfig)
• align_vision_model — AlignVisionConfig (modelo AlignVisionConfig)
• altclip — AltCLIPConfig (modelo AltCLIPConfig)
• altclip_text_model — AltCLIPTextConfig (modelo AltCLIPTextConfig)
• altclip_vision_model — AltCLIPVisionConfig (modelo AltCLIPVisionConfig)
• apertus — ApertusConfig (modelo ApertusConfig)
• arcee — ArceeConfig (modelo ArceeConfig)
• aria — AriaConfig (modelo AriaConfig)
• aria_text — AriaTextConfig (modelo AriaTextConfig)
• audio-spectrogram-transformer — ASTConfig (modelo ASTConfig)
• audioflamingo3 — AudioFlamingo3Config (modelo AudioFlamingo3Config)
• audioflamingo3_encoder — AudioFlamingo3EncoderConfig (modelo AudioFlamingo3EncoderConfig)
• autoformer — AutoformerConfig (modelo AutoformerConfig)
• axk1 — AXK1Config (modelo AXK1Config)
• axk2 — AXK2Config (modelo AXK2Config)
• aya_vision — AyaVisionConfig (modelo AyaVisionConfig)
• bailing_hybrid — BailingHybridConfig (modelo BailingHybridConfig)
• bamba — BambaConfig (modelo BambaConfig)
• bark — BarkConfig (modelo BarkConfig)
• bart — BartConfig (modelo BartConfig)
• beam — BeamConfig (modelo BeamConfig)
• beit — BeitConfig (modelo BeitConfig)
• bert — BertConfig (modelo BertConfig)
• bert-generation — BertGenerationConfig (modelo BertGenerationConfig)
• big_bird — BigBirdConfig (modelo BigBirdConfig)
• bigbird_pegasus — BigBirdPegasusConfig (modelo BigBirdPegasusConfig)
• biogpt — BioGptConfig (modelo BioGptConfig)
• bit — BitConfig (modelo BitConfig)
• bitnet — BitNetConfig (modelo BitNetConfig)
• blenderbot — BlenderbotConfig (modelo BlenderbotConfig)
• blenderbot-small — BlenderbotSmallConfig (modelo BlenderbotSmallConfig)
• blip — BlipConfig (modelo BlipConfig)
• blip-2 — Blip2Config (modelo Blip2Config)
• blip_2_qformer — Blip2QFormerConfig (modelo Blip2QFormerConfig)
• blip_2_vision_model — Blip2VisionConfig (modelo Blip2VisionConfig)
• blip_text_model — BlipTextConfig (modelo BlipTextConfig)
• blip_vision_model — BlipVisionConfig (modelo BlipVisionConfig)
• bloom — BloomConfig (modelo BloomConfig)
• blt — BltConfig (modelo BltConfig)
• blt_global_transformer — BltGlobalTransformerConfig (modelo BltGlobalTransformerConfig)
• blt_local_decoder — BltLocalDecoderConfig (modelo BltLocalDecoderConfig)
• blt_local_encoder — BltLocalEncoderConfig (modelo BltLocalEncoderConfig)
• blt_patcher — BltPatcherConfig (modelo BltPatcherConfig)
• bridgetower — BridgeTowerConfig (modelo BridgeTowerConfig)
• bridgetower_text_model — BridgeTowerTextConfig (modelo BridgeTowerTextConfig)
• bridgetower_vision_model — BridgeTowerVisionConfig (modelo BridgeTowerVisionConfig)
• bros — BrosConfig (modelo BrosConfig)
• camembert — CamembertConfig (modelo CamembertConfig)
• canary — CanaryConfig (modelo CanaryConfig)
• canary_decoder — CanaryDecoderConfig (modelo CanaryDecoderConfig)
• canine — CanineConfig (modelo CanineConfig)
• chameleon — ChameleonConfig (modelo ChameleonConfig)
• chameleon_vqgan — ChameleonVQVAEConfig (modelo ChameleonVQVAEConfig)
• chinese_clip — ChineseCLIPConfig (modelo ChineseCLIPConfig)
• chinese_clip_text_model — ChineseCLIPTextConfig (modelo ChineseCLIPTextConfig)
• chinese_clip_vision_model — ChineseCLIPVisionConfig (modelo ChineseCLIPVisionConfig)
• chmv2 — CHMv2Config (modelo CHMv2Config)
• clap — ClapConfig (modelo ClapConfig)
• clap_audio_model — ClapAudioConfig (modelo ClapAudioConfig)
• clap_text_model — ClapTextConfig (modelo ClapTextConfig)
• clip — CLIPConfig (modelo CLIPConfig)
• clip_text_model — CLIPTextConfig (modelo CLIPTextConfig)
• clip_vision_model — CLIPVisionConfig (modelo CLIPVisionConfig)
• clipseg — CLIPSegConfig (modelo CLIPSegConfig)
• clipseg_text_model — CLIPSegTextConfig (modelo CLIPSegTextConfig)
• clipseg_vision_model — CLIPSegVisionConfig (modelo CLIPSegVisionConfig)
• clvp — ClvpConfig (modelo ClvpConfig)
• clvp_decoder — ClvpDecoderConfig (modelo ClvpDecoderConfig)
• clvp_encoder — ClvpEncoderConfig (modelo ClvpEncoderConfig)
• codegen — CodeGenConfig (modelo CodeGenConfig)
• cohere — CohereConfig (modelo CohereConfig)
• cohere2 — Cohere2Config (modelo Cohere2Config)
• cohere2_moe — Cohere2MoeConfig (modelo Cohere2MoeConfig)
• cohere2_vision — Cohere2VisionConfig (modelo Cohere2VisionConfig)
• cohere_asr — CohereAsrConfig (modelo CohereAsrConfig)
• cohere_compass — CohereCompassConfig (modelo CohereCompassConfig)
• cohere_compass_text — CohereCompassTextConfig (modelo CohereCompassTextConfig)
• cohere_compass_vision — CohereCompassVisionConfig (modelo CohereCompassVisionConfig)
• colmodernvbert — ColModernVBertConfig (modelo ColModernVBertConfig)
• colpali — ColPaliConfig (modelo ColPaliConfig)
• colqwen2 — ColQwen2Config (modelo ColQwen2Config)
• conditional_detr — ConditionalDetrConfig (modelo ConditionalDetrConfig)
• convbert — ConvBertConfig (modelo ConvBertConfig)
• convnext — ConvNextConfig (modelo ConvNextConfig)
• convnextv2 — ConvNextV2Config (modelo ConvNextV2Config)
• cosmos3_edge — Cosmos3EdgeConfig (modelo Cosmos3EdgeConfig)
• cosmos3_edge_text — Cosmos3EdgeTextConfig (modelo Cosmos3EdgeTextConfig)
• cosmos3_edge_vision — Cosmos3EdgeVisionConfig (modelo Cosmos3EdgeVisionConfig)
• cosmos3_omni — Cosmos3OmniConfig (modelo Cosmos3OmniConfig)
• cpmant — CpmAntConfig (modelo CpmAntConfig)
• csm — CsmConfig (modelo CsmConfig)
• csm_depth_decoder_model — CsmDepthDecoderConfig (modelo CsmDepthDecoderConfig)
• ctrl — CTRLConfig (modelo CTRLConfig)
• cvt — CvtConfig (modelo CvtConfig)
• cwm — CwmConfig (modelo CwmConfig)
• d_fine — DFineConfig (modelo DFineConfig)
• dab-detr — DabDetrConfig (modelo DabDetrConfig)
• dac — DacConfig (modelo DacConfig)
• data2vec-audio — Data2VecAudioConfig (modelo Data2VecAudioConfig)
• data2vec-text — Data2VecTextConfig (modelo Data2VecTextConfig)
• data2vec-vision — Data2VecVisionConfig (modelo Data2VecVisionConfig)
• dbrx — DbrxConfig (modelo DbrxConfig)
• deberta — DebertaConfig (modelo DebertaConfig)
• deberta-v2 — DebertaV2Config (modelo DebertaV2Config)
• decision_transformer — DecisionTransformerConfig (modelo DecisionTransformerConfig)
• deepseek_ocr2 — DeepseekOcr2Config (modelo DeepseekOcr2Config)
• deepseek_ocr2_encoder — DeepseekOcr2VisionEncoderConfig (modelo DeepseekOcr2VisionEncoderConfig)
• deepseek_ocr2_sam_vision_model — DeepseekOcr2SamVisionConfig (modelo DeepseekOcr2SamVisionConfig)
• deepseek_ocr2_text — DeepseekOcr2TextConfig (modelo DeepseekOcr2TextConfig)
• deepseek_ocr2_vision — DeepseekOcr2VisionConfig (modelo DeepseekOcr2VisionConfig)
• deepseek_v2 — DeepseekV2Config (modelo DeepseekV2Config)
• deepseek_v3 — DeepseekV3Config (modelo DeepseekV3Config)
• deepseek_v32 — DeepseekV32Config (modelo DeepseekV32Config)
• deepseek_v4 — DeepseekV4Config (modelo DeepseekV4Config)
• deepseek_vl — DeepseekVLConfig (modelo DeepseekVLConfig)
• deepseek_vl_hybrid — DeepseekVLHybridConfig (modelo DeepseekVLHybridConfig)
• deformable_detr — DeformableDetrConfig (modelo DeformableDetrConfig)
• deimv2 — Deimv2Config (modelo Deimv2Config)
• deit — DeiTConfig (modelo DeiTConfig)
• depth_anything — DepthAnythingConfig (modelo DepthAnythingConfig)
• depth_pro — DepthProConfig (modelo DepthProConfig)
• detr — DetrConfig (modelo DetrConfig)
• dia — DiaConfig (modelo DiaConfig)
• dia_decoder — DiaDecoderConfig (modelo DiaDecoderConfig)
• dia_encoder — DiaEncoderConfig (modelo DiaEncoderConfig)
• diffllama — DiffLlamaConfig (modelo DiffLlamaConfig)
• diffusion_gemma — DiffusionGemmaConfig (modelo DiffusionGemmaConfig)
• diffusion_gemma_text — DiffusionGemmaTextConfig (modelo DiffusionGemmaTextConfig)
• dinat — DinatConfig (modelo DinatConfig)
• dinov2 — Dinov2Config (modelo Dinov2Config)
• dinov2_with_registers — Dinov2WithRegistersConfig (modelo Dinov2WithRegistersConfig)
• dinov3_convnext — DINOv3ConvNextConfig (modelo DINOv3ConvNextConfig)
• dinov3_vit — DINOv3ViTConfig (modelo DINOv3ViTConfig)
• distilbert — DistilBertConfig (modelo DistilBertConfig)
• doge — DogeConfig (modelo DogeConfig)
• donut-swin — DonutSwinConfig (modelo DonutSwinConfig)
• dots1 — Dots1Config (modelo Dots1Config)
• dpr — DPRConfig (modelo DPRConfig)
• dpt — DPTConfig (modelo DPTConfig)
• edgetam — EdgeTamConfig (modelo EdgeTamConfig)
• edgetam_mask_decoder — EdgeTamMaskDecoderConfig (modelo EdgeTamMaskDecoderConfig)
• edgetam_prompt_encoder — EdgeTamPromptEncoderConfig (modelo EdgeTamPromptEncoderConfig)
• edgetam_video — EdgeTamVideoConfig (modelo EdgeTamVideoConfig)
• edgetam_video_mask_decoder — EdgeTamVideoMaskDecoderConfig (modelo EdgeTamVideoMaskDecoderConfig)
• edgetam_video_prompt_encoder — EdgeTamVideoPromptEncoderConfig (modelo EdgeTamVideoPromptEncoderConfig)
• edgetam_vision_model — EdgeTamVisionConfig (modelo EdgeTamVisionConfig)
• efficientloftr — EfficientLoFTRConfig (modelo EfficientLoFTRConfig)
• efficientnet — EfficientNetConfig (modelo EfficientNetConfig)
• electra — ElectraConfig (modelo ElectraConfig)
• embedding_gemma2 — EmbeddingGemma2Config (modelo EmbeddingGemma2Config)
• embedding_gemma2_text — EmbeddingGemma2TextConfig (modelo EmbeddingGemma2TextConfig)
• emu3 — Emu3Config (modelo Emu3Config)
• emu3_text_model — Emu3TextConfig (modelo Emu3TextConfig)
• emu3_vqgan — Emu3VQVAEConfig (modelo Emu3VQVAEConfig)
• encodec — EncodecConfig (modelo EncodecConfig)
• encoder-decoder — EncoderDecoderConfig (modelo EncoderDecoderConfig)
• eomt — EomtConfig (modelo EomtConfig)
• eomt_dinov3 — EomtDinov3Config (modelo EomtDinov3Config)
• ernie — ErnieConfig (modelo ErnieConfig)
• ernie4_5 — Ernie4_5Config (modelo Ernie4_5Config)
• ernie4_5_moe — Ernie4_5_MoeConfig (modelo Ernie4_5_MoeConfig)
• ernie4_5_vl_moe — Ernie4_5_VLMoeConfig (modelo Ernie4_5_VLMoeConfig)
• ernie4_5_vl_moe_text — Ernie4_5_VLMoeTextConfig (modelo Ernie4_5_VLMoeTextConfig)
• ernie4_5_vl_moe_vision — Ernie4_5_VLMoeVisionConfig (modelo Ernie4_5_VLMoeVisionConfig)
• esm — EsmConfig (modelo EsmConfig)
• esmc — EsmcConfig (modelo EsmcConfig)
• esmfold2 — EsmFold2Config (modelo EsmFold2Config)
• eurobert — EuroBertConfig (modelo EuroBertConfig)
• evolla — EvollaConfig (modelo EvollaConfig)
• exaone4 — Exaone4Config (modelo Exaone4Config)
• exaone4_5 — Exaone4_5_Config (modelo Exaone4_5_Config)
• exaone4_5_vision — Exaone4_5_VisionConfig (modelo Exaone4_5_VisionConfig)
• exaone_moe — ExaoneMoeConfig (modelo ExaoneMoeConfig)
• falcon — FalconConfig (modelo FalconConfig)
• falcon_h1 — FalconH1Config (modelo FalconH1Config)
• falcon_mamba — FalconMambaConfig (modelo FalconMambaConfig)
• fast_vlm — FastVlmConfig (modelo FastVlmConfig)
• fastspeech2_conformer — FastSpeech2ConformerConfig (modelo FastSpeech2ConformerConfig)
• fastspeech2_conformer_hifigan — FastSpeech2ConformerHifiGanConfig (modelo FastSpeech2ConformerHifiGanConfig)
• fastspeech2_conformer_with_hifigan — FastSpeech2ConformerWithHifiGanConfig (modelo FastSpeech2ConformerWithHifiGanConfig)
• flaubert — FlaubertConfig (modelo FlaubertConfig)
• flava — FlavaConfig (modelo FlavaConfig)
• flava_image_codebook — FlavaImageCodebookConfig (modelo FlavaImageCodebookConfig)
• flava_image_model — FlavaImageConfig (modelo FlavaImageConfig)
• flava_multimodal_model — FlavaMultimodalConfig (modelo FlavaMultimodalConfig)
• flava_text_model — FlavaTextConfig (modelo FlavaTextConfig)
• flex_olmo — FlexOlmoConfig (modelo FlexOlmoConfig)
• flint — FlintConfig (modelo FlintConfig)
• florence2 — Florence2Config (modelo Florence2Config)
• florence_vision — Florence2VisionConfig (modelo Florence2VisionConfig)
• fnet — FNetConfig (modelo FNetConfig)
• focalnet — FocalNetConfig (modelo FocalNetConfig)
• fsmt — FSMTConfig (modelo FSMTConfig)
• fun_asr_nano — FunAsrNanoConfig (modelo FunAsrNanoConfig)
• fun_asr_nano_encoder — FunAsrNanoEncoderConfig (modelo FunAsrNanoEncoderConfig)
• funnel — FunnelConfig (modelo FunnelConfig)
• fuyu — FuyuConfig (modelo FuyuConfig)
• gemma — GemmaConfig (modelo GemmaConfig)
• gemma2 — Gemma2Config (modelo Gemma2Config)
• gemma3 — Gemma3Config (modelo Gemma3Config)
• gemma3_text — Gemma3TextConfig (modelo Gemma3TextConfig)
• gemma3n — Gemma3nConfig (modelo Gemma3nConfig)
• gemma3n_audio — Gemma3nAudioConfig (modelo Gemma3nAudioConfig)
• gemma3n_text — Gemma3nTextConfig (modelo Gemma3nTextConfig)
• gemma3n_vision — Gemma3nVisionConfig (modelo Gemma3nVisionConfig)
• gemma4 — Gemma4Config (modelo Gemma4Config)
• gemma4_assistant — Gemma4AssistantConfig (modelo Gemma4AssistantConfig)
• gemma4_audio — Gemma4AudioConfig (modelo Gemma4AudioConfig)
• gemma4_text — Gemma4TextConfig (modelo Gemma4TextConfig)
• gemma4_unified — Gemma4UnifiedConfig (modelo Gemma4UnifiedConfig)
• gemma4_unified_assistant — Gemma4UnifiedAssistantConfig (modelo Gemma4UnifiedAssistantConfig)
• gemma4_unified_audio — Gemma4UnifiedAudioConfig (modelo Gemma4UnifiedAudioConfig)
• gemma4_unified_text — Gemma4UnifiedTextConfig (modelo Gemma4UnifiedTextConfig)
• gemma4_unified_vision — Gemma4UnifiedVisionConfig (modelo Gemma4UnifiedVisionConfig)
• gemma4_vision — Gemma4VisionConfig (modelo Gemma4VisionConfig)
• git — GitConfig (modelo GitConfig)
• git_vision_model — GitVisionConfig (modelo GitVisionConfig)
• glm — GlmConfig (modelo GlmConfig)
• glm4 — Glm4Config (modelo Glm4Config)
• glm46v — Glm46VConfig (modelo Glm46VConfig)
• glm4_moe — Glm4MoeConfig (modelo Glm4MoeConfig)
• glm4_moe_lite — Glm4MoeLiteConfig (modelo Glm4MoeLiteConfig)
• glm4v — Glm4vConfig (modelo Glm4vConfig)
• glm4v_moe — Glm4vMoeConfig (modelo Glm4vMoeConfig)
• glm4v_moe_text — Glm4vMoeTextConfig (modelo Glm4vMoeTextConfig)
• glm4v_moe_vision — Glm4vMoeVisionConfig (modelo Glm4vMoeVisionConfig)
• glm4v_text — Glm4vTextConfig (modelo Glm4vTextConfig)
• glm4v_vision — Glm4vVisionConfig (modelo Glm4vVisionConfig)
• glm5_next — Glm5NextConfig (modelo Glm5NextConfig)
• glm5_next_text — Glm5NextTextConfig (modelo Glm5NextTextConfig)
• glm5_next_vision — Glm5NextVisionConfig (modelo Glm5NextVisionConfig)
• glm_image — GlmImageConfig (modelo GlmImageConfig)
• glm_image_text — GlmImageTextConfig (modelo GlmImageTextConfig)
• glm_image_vision — GlmImageVisionConfig (modelo GlmImageVisionConfig)
• glm_image_vqmodel — GlmImageVQVAEConfig (modelo GlmImageVQVAEConfig)
• glm_moe_dsa — GlmMoeDsaConfig (modelo GlmMoeDsaConfig)
• glm_ocr — GlmOcrConfig (modelo GlmOcrConfig)
• glm_ocr_text — GlmOcrTextConfig (modelo GlmOcrTextConfig)
• glm_ocr_vision — GlmOcrVisionConfig (modelo GlmOcrVisionConfig)
• glmasr — GlmAsrConfig (modelo GlmAsrConfig)
• glmasr_encoder — GlmAsrEncoderConfig (modelo GlmAsrEncoderConfig)
• glmga — GlmgaConfig (modelo GlmgaConfig)
• glpn — GLPNConfig (modelo GLPNConfig)
• got_ocr2 — GotOcr2Config (modelo GotOcr2Config)
• got_ocr2_vision — GotOcr2VisionConfig (modelo GotOcr2VisionConfig)
• gpt-sw3 — GPT2Config (modelo GPT2Config)
• gpt2 — GPT2Config (modelo GPT2Config)
• gpt_bigcode — GPTBigCodeConfig (modelo GPTBigCodeConfig)
• gpt_neo — GPTNeoConfig (modelo GPTNeoConfig)
• gpt_neox — GPTNeoXConfig (modelo GPTNeoXConfig)
• gpt_neox_japanese — GPTNeoXJapaneseConfig (modelo GPTNeoXJapaneseConfig)
• gpt_oss — GptOssConfig (modelo GptOssConfig)
• gptj — GPTJConfig (modelo GPTJConfig)
• gradium_phonon_tts — GradiumPhononTTSConfig (modelo GradiumPhononTTSConfig)
• granite — GraniteConfig (modelo GraniteConfig)
• granite4_vision — Granite4VisionConfig (modelo Granite4VisionConfig)
• granite4_vision_text — Granite4VisionTextConfig (modelo Granite4VisionTextConfig)
• granite_speech — GraniteSpeechConfig (modelo GraniteSpeechConfig)
• granite_speech5_ctc — GraniteSpeech5CTCConfig (modelo GraniteSpeech5CTCConfig)
• granite_speech5_encoder — GraniteSpeech5EncoderConfig (modelo GraniteSpeech5EncoderConfig)
• granite_speech_encoder — GraniteSpeechEncoderConfig (modelo GraniteSpeechEncoderConfig)
• granite_speech_plus — GraniteSpeechPlusConfig (modelo GraniteSpeechPlusConfig)
• granite_speech_plus_encoder — GraniteSpeechPlusEncoderConfig (modelo GraniteSpeechPlusEncoderConfig)
• granite_swa — GraniteSWAConfig (modelo GraniteSWAConfig)
• granitemoe — GraniteMoeConfig (modelo GraniteMoeConfig)
• granitemoe_swa — GraniteMoeSWAConfig (modelo GraniteMoeSWAConfig)
• granitemoehybrid — GraniteMoeHybridConfig (modelo GraniteMoeHybridConfig)
• granitemoeshared — GraniteMoeSharedConfig (modelo GraniteMoeSharedConfig)
• grounding-dino — GroundingDinoConfig (modelo GroundingDinoConfig)
• groupvit — GroupViTConfig (modelo GroupViTConfig)
• groupvit_text_model — GroupViTTextConfig (modelo GroupViTTextConfig)
• groupvit_vision_model — GroupViTVisionConfig (modelo GroupViTVisionConfig)
• gte — GteConfig (modelo GteConfig)
• helium — HeliumConfig (modelo HeliumConfig)
• hgnet_v2 — HGNetV2Config (modelo HGNetV2Config)
• hiera — HieraConfig (modelo HieraConfig)
• higgs_audio_v2 — HiggsAudioV2Config (modelo HiggsAudioV2Config)
• higgs_audio_v2_tokenizer — HiggsAudioV2TokenizerConfig (modelo HiggsAudioV2TokenizerConfig)
• hrm_text — HrmTextConfig (modelo HrmTextConfig)
• hubert — HubertConfig (modelo HubertConfig)
• hunyuan_v1_dense — HunYuanDenseV1Config (modelo HunYuanDenseV1Config)
• hunyuan_v1_moe — HunYuanMoEV1Config (modelo HunYuanMoEV1Config)
• hunyuan_vl — HunYuanVLConfig (modelo HunYuanVLConfig)
• hunyuan_vl_text — HunYuanVLTextConfig (modelo HunYuanVLTextConfig)
• hunyuan_vl_vision — HunYuanVLVisionConfig (modelo HunYuanVLVisionConfig)
• hy_v3 — HYV3Config (modelo HYV3Config)
• hy_v4 — HYV4Config (modelo HYV4Config)
• hyperclovax — HyperCLOVAXConfig (modelo HyperCLOVAXConfig)
• hyperclovax_vision_v2 — HyperCLOVAXVisionV2Config (modelo HyperCLOVAXVisionV2Config)
• ibert — IBertConfig (modelo IBertConfig)
• idefics — IdeficsConfig (modelo IdeficsConfig)
• idefics2 — Idefics2Config (modelo Idefics2Config)
• idefics2_perceiver — Idefics2PerceiverConfig (modelo Idefics2PerceiverConfig)
• idefics2_vision — Idefics2VisionConfig (modelo Idefics2VisionConfig)
• idefics3 — Idefics3Config (modelo Idefics3Config)
• idefics3_vision — Idefics3VisionConfig (modelo Idefics3VisionConfig)
• idefics_perciever — IdeficsPerceiverConfig (modelo IdeficsPerceiverConfig)
• idefics_vision — IdeficsVisionConfig (modelo IdeficsVisionConfig)
• ijepa — IJepaConfig (modelo IJepaConfig)
• imagegpt — ImageGPTConfig (modelo ImageGPTConfig)
• informer — InformerConfig (modelo InformerConfig)
• inkling_audio — InklingAudioConfig (modelo InklingAudioConfig)
• inkling_mm_model — InklingConfig (modelo InklingConfig)
• inkling_text — InklingTextConfig (modelo InklingTextConfig)
• inkling_vision — InklingVisionConfig (modelo InklingVisionConfig)
• instructblip — InstructBlipConfig (modelo InstructBlipConfig)
• instructblip_qformer — InstructBlipQFormerConfig (modelo InstructBlipQFormerConfig)
• instructblip_vision_model — InstructBlipVisionConfig (modelo InstructBlipVisionConfig)
• instructblipvideo — InstructBlipVideoConfig (modelo InstructBlipVideoConfig)
• instructblipvideo_qformer — InstructBlipVideoQFormerConfig (modelo InstructBlipVideoQFormerConfig)
• instructblipvideo_vision_model — InstructBlipVideoVisionConfig (modelo InstructBlipVideoVisionConfig)
• internvl — InternVLConfig (modelo InternVLConfig)
• internvl_vision — InternVLVisionConfig (modelo InternVLVisionConfig)
• jais2 — Jais2Config (modelo Jais2Config)
• jamba — JambaConfig (modelo JambaConfig)
• janus — JanusConfig (modelo JanusConfig)
• janus_vision_model — JanusVisionConfig (modelo JanusVisionConfig)
• janus_vqgan — JanusVQVAEConfig (modelo JanusVQVAEConfig)
• jetmoe — JetMoeConfig (modelo JetMoeConfig)
• jina_embeddings_v3 — JinaEmbeddingsV3Config (modelo JinaEmbeddingsV3Config)
• kimi_k25 — Kimi_K25Config (modelo Kimi_K25Config)
• kimi_k25_vision — Kimi_K25VisionConfig (modelo Kimi_K25VisionConfig)
• kimi_linear — KimiLinearConfig (modelo KimiLinearConfig)
• kolibri1 — Kolibri1Config (modelo Kolibri1Config)
• kosmos-2 — Kosmos2Config (modelo Kosmos2Config)
• kosmos-2.5 — Kosmos2_5Config (modelo Kosmos2_5Config)
• kosmos_2_5_text_model — Kosmos2_5TextConfig (modelo Kosmos2_5TextConfig)
• kosmos_2_5_vision_model — Kosmos2_5VisionConfig (modelo Kosmos2_5VisionConfig)
• kosmos_2_text_model — Kosmos2TextConfig (modelo Kosmos2TextConfig)
• kosmos_2_vision_model — Kosmos2VisionConfig (modelo Kosmos2VisionConfig)
• kyutai_speech_to_text — KyutaiSpeechToTextConfig (modelo KyutaiSpeechToTextConfig)
• laguna — LagunaConfig (modelo LagunaConfig)
• lasr_ctc — LasrCTCConfig (modelo LasrCTCConfig)
• lasr_encoder — LasrEncoderConfig (modelo LasrEncoderConfig)
• layoutlm — LayoutLMConfig (modelo LayoutLMConfig)
• layoutlmv2 — LayoutLMv2Config (modelo LayoutLMv2Config)
• layoutlmv3 — LayoutLMv3Config (modelo LayoutLMv3Config)
• layoutxlm — LayoutXLMConfig (modelo LayoutXLMConfig)
• led — LEDConfig (modelo LEDConfig)
• levit — LevitConfig (modelo LevitConfig)
• lfm2 — Lfm2Config (modelo Lfm2Config)
• lfm2_moe — Lfm2MoeConfig (modelo Lfm2MoeConfig)
• lfm2_vl — Lfm2VlConfig (modelo Lfm2VlConfig)
• lightglue — LightGlueConfig (modelo LightGlueConfig)
• lighton_ocr — LightOnOcrConfig (modelo LightOnOcrConfig)
• lilt — LiltConfig (modelo LiltConfig)
• llama — LlamaConfig (modelo LlamaConfig)
• llama4 — Llama4Config (modelo Llama4Config)
• llama4_text — Llama4TextConfig (modelo Llama4TextConfig)
• llama4_vision_model — Llama4VisionConfig (modelo Llama4VisionConfig)
• llava — LlavaConfig (modelo LlavaConfig)
• llava_next — LlavaNextConfig (modelo LlavaNextConfig)
• llava_next_video — LlavaNextVideoConfig (modelo LlavaNextVideoConfig)
• llava_onevision — LlavaOnevisionConfig (modelo LlavaOnevisionConfig)
• longcat_flash — LongcatFlashConfig (modelo LongcatFlashConfig)
• longformer — LongformerConfig (modelo LongformerConfig)
• longt5 — LongT5Config (modelo LongT5Config)
• luke — LukeConfig (modelo LukeConfig)
• lw_detr — LwDetrConfig (modelo LwDetrConfig)
• lw_detr_vit — LwDetrViTConfig (modelo LwDetrViTConfig)
• lxmert — LxmertConfig (modelo LxmertConfig)
• m2m_100 — M2M100Config (modelo M2M100Config)
• mamba — MambaConfig (modelo MambaConfig)
• mamba2 — Mamba2Config (modelo Mamba2Config)
• marian — MarianConfig (modelo MarianConfig)
• markuplm — MarkupLMConfig (modelo MarkupLMConfig)
• mask2former — Mask2FormerConfig (modelo Mask2FormerConfig)
• maskformer — MaskFormerConfig (modelo MaskFormerConfig)
• maskformer-swin — MaskFormerSwinConfig (modelo MaskFormerSwinConfig)
• mbart — MBartConfig (modelo MBartConfig)
• megatron-bert — MegatronBertConfig (modelo MegatronBertConfig)
• mellum — MellumConfig (modelo MellumConfig)
• metaclip_2 — MetaClip2Config (modelo MetaClip2Config)
• metaclip_2_text_model — MetaClip2TextConfig (modelo MetaClip2TextConfig)
• metaclip_2_vision_model — MetaClip2VisionConfig (modelo MetaClip2VisionConfig)
• mgp-str — MgpstrConfig (modelo MgpstrConfig)
• mimi — MimiConfig (modelo MimiConfig)
• mimi_decoder — MimiDecoderConfig (modelo MimiDecoderConfig)
• mimi_encoder — MimiEncoderConfig (modelo MimiEncoderConfig)
• mimo_v2_flash — MiMoV2FlashConfig (modelo MiMoV2FlashConfig)
• minicpm3 — MiniCPM3Config (modelo MiniCPM3Config)
• minicpm_ocr — MiniCPMOcrConfig (modelo MiniCPMOcrConfig)
• minicpm_ocr_vision — MiniCPMOcrVisionConfig (modelo MiniCPMOcrVisionConfig)
• minicpmv4_6 — MiniCPMV4_6Config (modelo MiniCPMV4_6Config)
• minicpmv4_6_vision — MiniCPMV4_6VisionConfig (modelo MiniCPMV4_6VisionConfig)
• minicpmv4_7 — MiniCPMV4_7Config (modelo MiniCPMV4_7Config)
• minicpmv4_7_vision — MiniCPMV4_7VisionConfig (modelo MiniCPMV4_7VisionConfig)
• minimax — MiniMaxConfig (modelo MiniMaxConfig)
• minimax_m2 — MiniMaxM2Config (modelo MiniMaxM2Config)
• minimax_m3_vl — MiniMaxM3VLConfig (modelo MiniMaxM3VLConfig)
• minimax_m3_vl_text — MiniMaxM3VLTextConfig (modelo MiniMaxM3VLTextConfig)
• minimax_m3_vl_vision — MiniMaxM3VLVisionConfig (modelo MiniMaxM3VLVisionConfig)
• ministral — MinistralConfig (modelo MinistralConfig)
• ministral3 — Ministral3Config (modelo Ministral3Config)
• mistral — MistralConfig (modelo MistralConfig)
• mistral3 — Mistral3Config (modelo Mistral3Config)
• mistral4 — Mistral4Config (modelo Mistral4Config)
• mistral_large4 — MistralLarge4Config (modelo MistralLarge4Config)
• mistral_large4_text — MistralLarge4TextConfig (modelo MistralLarge4TextConfig)
• mixtral — MixtralConfig (modelo MixtralConfig)
• mlcd — MLCDVisionConfig (modelo MLCDVisionConfig)
• mlcd_vision_model — MLCDVisionConfig (modelo MLCDVisionConfig)
• mllama — MllamaConfig (modelo MllamaConfig)
• mllama_text_model — MllamaTextConfig (modelo MllamaTextConfig)
• mllama_vision_model — MllamaVisionConfig (modelo MllamaVisionConfig)
• mm-grounding-dino — MMGroundingDinoConfig (modelo MMGroundingDinoConfig)
• mobilebert — MobileBertConfig (modelo MobileBertConfig)
• mobilenet_v1 — MobileNetV1Config (modelo MobileNetV1Config)
• mobilenet_v2 — MobileNetV2Config (modelo MobileNetV2Config)
• mobilevit — MobileViTConfig (modelo MobileViTConfig)
• mobilevitv2 — MobileViTV2Config (modelo MobileViTV2Config)
• modernbert — ModernBertConfig (modelo ModernBertConfig)
• modernbert-decoder — ModernBertDecoderConfig (modelo ModernBertDecoderConfig)
• modernvbert — ModernVBertConfig (modelo ModernVBertConfig)
• molmo2 — Molmo2Config (modelo Molmo2Config)
• molmo2_adapter — Molmo2AdapterConfig (modelo Molmo2AdapterConfig)
• molmo2_text — Molmo2TextConfig (modelo Molmo2TextConfig)
• molmo2_vision — Molmo2VisionConfig (modelo Molmo2VisionConfig)
• moonshine — MoonshineConfig (modelo MoonshineConfig)
• moonshine_streaming — MoonshineStreamingConfig (modelo MoonshineStreamingConfig)
• moonshine_streaming_encoder — MoonshineStreamingEncoderConfig (modelo MoonshineStreamingEncoderConfig)
• moshi — MoshiConfig (modelo MoshiConfig)
• moshi_depth — MoshiDepthConfig (modelo MoshiDepthConfig)
• moss_transcribe_diarize — MossTranscribeDiarizeConfig (modelo MossTranscribeDiarizeConfig)
• moss_transcribe_diarize_audio — MossTranscribeDiarizeAudioConfig (modelo MossTranscribeDiarizeAudioConfig)
• mpnet — MPNetConfig (modelo MPNetConfig)
• mpt — MptConfig (modelo MptConfig)
• mra — MraConfig (modelo MraConfig)
• mt5 — MT5Config (modelo MT5Config)
• muse_glimmer — MuseGlimmerConfig (modelo MuseGlimmerConfig)
• muse_glimmer_assistant — MuseGlimmerAssistantConfig (modelo MuseGlimmerAssistantConfig)
• muse_glimmer_text — MuseGlimmerTextConfig (modelo MuseGlimmerTextConfig)
• muse_glimmer_vision — MuseGlimmerVisionConfig (modelo MuseGlimmerVisionConfig)
• muse_spark — MuseSparkConfig (modelo MuseSparkConfig)
• muse_spark_audio — MuseSparkAudioConfig (modelo MuseSparkAudioConfig)
• muse_spark_text — MuseSparkTextConfig (modelo MuseSparkTextConfig)
• muse_spark_vision — MuseSparkVisionConfig (modelo MuseSparkVisionConfig)
• musicflamingo — MusicFlamingoConfig (modelo MusicFlamingoConfig)
• musicgen — MusicgenConfig (modelo MusicgenConfig)
• musicgen_decoder — MusicgenDecoderConfig (modelo MusicgenDecoderConfig)
• musicgen_melody — MusicgenMelodyConfig (modelo MusicgenMelodyConfig)
• musicgen_melody_decoder — MusicgenMelodyDecoderConfig (modelo MusicgenMelodyDecoderConfig)
• mvp — MvpConfig (modelo MvpConfig)
• nanochat — NanoChatConfig (modelo NanoChatConfig)
• nemo_speechlm — Nemotron3_5TranscribeConfig (modelo Nemotron3_5TranscribeConfig)
• nemotron — NemotronConfig (modelo NemotronConfig)
• nemotron3_5_asr — Nemotron3_5AsrConfig (modelo Nemotron3_5AsrConfig)
• nemotron3_5_transcribe_asr_audio — Nemotron3_5TranscribeAsrAudioConfig (modelo Nemotron3_5TranscribeAsrAudioConfig)
• nemotron3_5_transcribe_audio — Nemotron3_5TranscribeAudioConfig (modelo Nemotron3_5TranscribeAudioConfig)
• nemotron3_5_transcribe_word_timestamp — Nemotron3_5TranscribeWordTimestampConfig (modelo Nemotron3_5TranscribeWordTimestampConfig)
• nemotron3_diarization — Nemotron3DiarizationConfig (modelo Nemotron3DiarizationConfig)
• nemotron3_diarization_audio — Nemotron3DiarizationAudioConfig (modelo Nemotron3DiarizationAudioConfig)
• nemotron_asr_streaming — NemotronAsrStreamingConfig (modelo NemotronAsrStreamingConfig)
• nemotron_asr_streaming_encoder — NemotronAsrStreamingEncoderConfig (modelo NemotronAsrStreamingEncoderConfig)
• nemotron_h — NemotronHConfig (modelo NemotronHConfig)
• nemotron_h_omni — NemotronH_Omni_Reasoning_V3_Config (modelo NemotronH_Omni_Reasoning_V3_Config)
• neomme — NeoMMEConfig (modelo NeoMMEConfig)
• neucodec — NeuCodecConfig (modelo NeuCodecConfig)
• nllb-moe — NllbMoeConfig (modelo NllbMoeConfig)
• nomic_bert — NomicBertConfig (modelo NomicBertConfig)
• nougat — NougatConfig (modelo NougatConfig)
• nystromformer — NystromformerConfig (modelo NystromformerConfig)
• olmo — OlmoConfig (modelo OlmoConfig)
• olmo2 — Olmo2Config (modelo Olmo2Config)
• olmo3 — Olmo3Config (modelo Olmo3Config)
• olmo_hybrid — OlmoHybridConfig (modelo OlmoHybridConfig)
• olmoe — OlmoeConfig (modelo OlmoeConfig)
• omdet-turbo — OmDetTurboConfig (modelo OmDetTurboConfig)
• oneformer — OneFormerConfig (modelo OneFormerConfig)
• openai-gpt — OpenAIGPTConfig (modelo OpenAIGPTConfig)
• openai_privacy_filter — OpenAIPrivacyFilterConfig (modelo OpenAIPrivacyFilterConfig)
• opt — OPTConfig (modelo OPTConfig)
• ovis2 — Ovis2Config (modelo Ovis2Config)
• ovis2_vision — Ovis2VisionConfig (modelo Ovis2VisionConfig)
• owlv2 — Owlv2Config (modelo Owlv2Config)
• owlv2_text_model — Owlv2TextConfig (modelo Owlv2TextConfig)
• owlv2_vision_model — Owlv2VisionConfig (modelo Owlv2VisionConfig)
• owlvit — OwlViTConfig (modelo OwlViTConfig)
• owlvit_text_model — OwlViTTextConfig (modelo OwlViTTextConfig)
• owlvit_vision_model — OwlViTVisionConfig (modelo OwlViTVisionConfig)
• paddleocr_vl — PaddleOCRVLConfig (modelo PaddleOCRVLConfig)
• paddleocr_vl_text — PaddleOCRTextConfig (modelo PaddleOCRTextConfig)
• paddleocr_vl_vision — PaddleOCRVisionConfig (modelo PaddleOCRVisionConfig)
• paligemma — PaliGemmaConfig (modelo PaliGemmaConfig)
• parakeet_ctc — ParakeetCTCConfig (modelo ParakeetCTCConfig)
• parakeet_encoder — ParakeetEncoderConfig (modelo ParakeetEncoderConfig)
• parakeet_rnnt — ParakeetRNNTConfig (modelo ParakeetRNNTConfig)
• parakeet_tdt — ParakeetTDTConfig (modelo ParakeetTDTConfig)
• patchtsmixer — PatchTSMixerConfig (modelo PatchTSMixerConfig)
• patchtst — PatchTSTConfig (modelo PatchTSTConfig)
• pe_audio — PeAudioConfig (modelo PeAudioConfig)
• pe_audio_encoder — PeAudioEncoderConfig (modelo PeAudioEncoderConfig)
• pe_audio_video — PeAudioVideoConfig (modelo PeAudioVideoConfig)
• pe_audio_video_encoder — PeAudioVideoEncoderConfig (modelo PeAudioVideoEncoderConfig)
• pe_video — PeVideoConfig (modelo PeVideoConfig)
• pe_video_encoder — PeVideoEncoderConfig (modelo PeVideoEncoderConfig)
• pegasus — PegasusConfig (modelo PegasusConfig)
• pegasus_x — PegasusXConfig (modelo PegasusXConfig)
• perceiver — PerceiverConfig (modelo PerceiverConfig)
• perception_lm — PerceptionLMConfig (modelo PerceptionLMConfig)
• persimmon — PersimmonConfig (modelo PersimmonConfig)
• phi — PhiConfig (modelo PhiConfig)
• phi3 — Phi3Config (modelo Phi3Config)
• phi4_multimodal — Phi4MultimodalConfig (modelo Phi4MultimodalConfig)
• phi4_multimodal_audio — Phi4MultimodalAudioConfig (modelo Phi4MultimodalAudioConfig)
• phi4_multimodal_vision — Phi4MultimodalVisionConfig (modelo Phi4MultimodalVisionConfig)
• phimoe — PhimoeConfig (modelo PhimoeConfig)
• pi0 — PI0Config (modelo PI0Config)
• pix2struct — Pix2StructConfig (modelo Pix2StructConfig)
• pix2struct_text_model — Pix2StructTextConfig (modelo Pix2StructTextConfig)
• pix2struct_vision_model — Pix2StructVisionConfig (modelo Pix2StructVisionConfig)
• pixio — PixioConfig (modelo PixioConfig)
• pixtral — PixtralVisionConfig (modelo PixtralVisionConfig)
• plbart — PLBartConfig (modelo PLBartConfig)
• pocket_tts — PocketTTSConfig (modelo PocketTTSConfig)
• pocket_tts_audio_tokenizer — PocketTTSAudioTokenizerConfig (modelo PocketTTSAudioTokenizerConfig)
• pocket_tts_audio_tokenizer_decoder — PocketTTSAudioTokenizerDecoderConfig (modelo PocketTTSAudioTokenizerDecoderConfig)
• pocket_tts_audio_tokenizer_encoder — PocketTTSAudioTokenizerEncoderConfig (modelo PocketTTSAudioTokenizerEncoderConfig)
• poolformer — PoolFormerConfig (modelo PoolFormerConfig)
• pop2piano — Pop2PianoConfig (modelo Pop2PianoConfig)
• pp_chart2table — PPChart2TableConfig (modelo PPChart2TableConfig)
• pp_chart2table_vision — PPChart2TableVisionConfig (modelo PPChart2TableVisionConfig)
• pp_doclayout_v2 — PPDocLayoutV2Config (modelo PPDocLayoutV2Config)
• pp_doclayout_v3 — PPDocLayoutV3Config (modelo PPDocLayoutV3Config)
• pp_doclayout_v4 — PPDocLayoutV4Config (modelo PPDocLayoutV4Config)
• pp_formulanet — PPFormulaNetConfig (modelo PPFormulaNetConfig)
• pp_formulanet_text — PPFormulaNetTextConfig (modelo PPFormulaNetTextConfig)
• pp_formulanet_vision — PPFormulaNetVisionConfig (modelo PPFormulaNetVisionConfig)
• pp_lcnet — PPLCNetConfig (modelo PPLCNetConfig)
• pp_lcnet_v3 — PPLCNetV3Config (modelo PPLCNetV3Config)
• pp_lcnet_v4 — PPLCNetV4Config (modelo PPLCNetV4Config)
• pp_ocrv5_mobile_det — PPOCRV5MobileDetConfig (modelo PPOCRV5MobileDetConfig)
• pp_ocrv5_mobile_rec — PPOCRV5MobileRecConfig (modelo PPOCRV5MobileRecConfig)
• pp_ocrv5_server_det — PPOCRV5ServerDetConfig (modelo PPOCRV5ServerDetConfig)
• pp_ocrv5_server_rec — PPOCRV5ServerRecConfig (modelo PPOCRV5ServerRecConfig)
• pp_ocrv6_medium_det — PPOCRV6MediumDetConfig (modelo PPOCRV6MediumDetConfig)
• pp_ocrv6_small_det — PPOCRV6SmallDetConfig (modelo PPOCRV6SmallDetConfig)
• pp_ocrv6_small_rec — PPOCRV6SmallRecConfig (modelo PPOCRV6SmallRecConfig)
• pp_ocrv6_tiny_rec — PPOCRV6TinyRecConfig (modelo PPOCRV6TinyRecConfig)
• prompt_depth_anything — PromptDepthAnythingConfig (modelo PromptDepthAnythingConfig)
• prophetnet — ProphetNetConfig (modelo ProphetNetConfig)
• pvt — PvtConfig (modelo PvtConfig)
• pvt_v2 — PvtV2Config (modelo PvtV2Config)
• qianfan_ocr — QianfanOCRConfig (modelo QianfanOCRConfig)
• qianfan_ocr_vision — QianfanOCRVisionConfig (modelo QianfanOCRVisionConfig)
• qwen2 — Qwen2Config (modelo Qwen2Config)
• qwen2_5_omni — Qwen2_5OmniConfig (modelo Qwen2_5OmniConfig)
• qwen2_5_omni_audio_encoder — Qwen2_5OmniAudioEncoderConfig (modelo Qwen2_5OmniAudioEncoderConfig)
• qwen2_5_omni_bigvgan — Qwen2_5OmniBigVGANConfig (modelo Qwen2_5OmniBigVGANConfig)
• qwen2_5_omni_dit — Qwen2_5OmniDiTConfig (modelo Qwen2_5OmniDiTConfig)
• qwen2_5_omni_talker — Qwen2_5OmniTalkerConfig (modelo Qwen2_5OmniTalkerConfig)
• qwen2_5_omni_text — Qwen2_5OmniTextConfig (modelo Qwen2_5OmniTextConfig)
• qwen2_5_omni_thinker — Qwen2_5OmniThinkerConfig (modelo Qwen2_5OmniThinkerConfig)
• qwen2_5_omni_token2wav — Qwen2_5OmniToken2WavConfig (modelo Qwen2_5OmniToken2WavConfig)
• qwen2_5_omni_vision_encoder — Qwen2_5OmniVisionEncoderConfig (modelo Qwen2_5OmniVisionEncoderConfig)
• qwen2_5_vl — Qwen2_5_VLConfig (modelo Qwen2_5_VLConfig)
• qwen2_5_vl_text — Qwen2_5_VLTextConfig (modelo Qwen2_5_VLTextConfig)
• qwen2_5_vl_vision — Qwen2_5_VLVisionConfig (modelo Qwen2_5_VLVisionConfig)
• qwen2_audio — Qwen2AudioConfig (modelo Qwen2AudioConfig)
• qwen2_audio_encoder — Qwen2AudioEncoderConfig (modelo Qwen2AudioEncoderConfig)
• qwen2_moe — Qwen2MoeConfig (modelo Qwen2MoeConfig)
• qwen2_vl — Qwen2VLConfig (modelo Qwen2VLConfig)
• qwen2_vl_text — Qwen2VLTextConfig (modelo Qwen2VLTextConfig)
• qwen2_vl_vision — Qwen2VLVisionConfig (modelo Qwen2VLVisionConfig)
• qwen3 — Qwen3Config (modelo Qwen3Config)
• qwen3_5 — Qwen3_5Config (modelo Qwen3_5Config)
• qwen3_5_moe — Qwen3_5MoeConfig (modelo Qwen3_5MoeConfig)
• qwen3_5_moe_text — Qwen3_5MoeTextConfig (modelo Qwen3_5MoeTextConfig)
• qwen3_5_moe_vision — Qwen3_5MoeVisionConfig (modelo Qwen3_5MoeVisionConfig)
• qwen3_5_text — Qwen3_5TextConfig (modelo Qwen3_5TextConfig)
• qwen3_5_vision — Qwen3_5VisionConfig (modelo Qwen3_5VisionConfig)
• qwen3_asr — Qwen3ASRConfig (modelo Qwen3ASRConfig)
• qwen3_asr_encoder — Qwen3ASREncoderConfig (modelo Qwen3ASREncoderConfig)
• qwen3_moe — Qwen3MoeConfig (modelo Qwen3MoeConfig)
• qwen3_next — Qwen3NextConfig (modelo Qwen3NextConfig)
• qwen3_omni_moe — Qwen3OmniMoeConfig (modelo Qwen3OmniMoeConfig)
• qwen3_omni_moe_audio_encoder — Qwen3OmniMoeAudioEncoderConfig (modelo Qwen3OmniMoeAudioEncoderConfig)
• qwen3_omni_moe_talker_code_predictor — Qwen3OmniMoeTalkerCodePredictorConfig (modelo Qwen3OmniMoeTalkerCodePredictorConfig)
• qwen3_omni_moe_talker_text — Qwen3OmniMoeTalkerTextConfig (modelo Qwen3OmniMoeTalkerTextConfig)
• qwen3_omni_moe_text — Qwen3OmniMoeTextConfig (modelo Qwen3OmniMoeTextConfig)
• qwen3_omni_moe_thinker — Qwen3OmniMoeThinkerConfig (modelo Qwen3OmniMoeThinkerConfig)
• qwen3_omni_moe_vision_encoder — Qwen3OmniMoeVisionEncoderConfig (modelo Qwen3OmniMoeVisionEncoderConfig)
• qwen3_vl — Qwen3VLConfig (modelo Qwen3VLConfig)
• qwen3_vl_moe — Qwen3VLMoeConfig (modelo Qwen3VLMoeConfig)
• qwen3_vl_moe_text — Qwen3VLMoeTextConfig (modelo Qwen3VLMoeTextConfig)
• qwen3_vl_moe_vision — Qwen3VLMoeVisionConfig (modelo Qwen3VLMoeVisionConfig)
• qwen3_vl_text — Qwen3VLTextConfig (modelo Qwen3VLTextConfig)
• qwen3_vl_vision — Qwen3VLVisionConfig (modelo Qwen3VLVisionConfig)
• qwen4_exp — Qwen4ExpConfig (modelo Qwen4ExpConfig)
• qwen4_exp_text — Qwen4ExpTextConfig (modelo Qwen4ExpTextConfig)
• qwen4_exp_vision — Qwen4ExpVisionConfig (modelo Qwen4ExpVisionConfig)
• radio — RadioConfig (modelo RadioConfig)
• rag — RagConfig (modelo RagConfig)
• recurrent_gemma — RecurrentGemmaConfig (modelo RecurrentGemmaConfig)
• reformer — ReformerConfig (modelo ReformerConfig)
• regnet — RegNetConfig (modelo RegNetConfig)
• rembert — RemBertConfig (modelo RemBertConfig)
• resnet — ResNetConfig (modelo ResNetConfig)
• rf_detr — RfDetrConfig (modelo RfDetrConfig)
• rf_detr_dinov2 — RfDetrDinov2Config (modelo RfDetrDinov2Config)
• roberta — RobertaConfig (modelo RobertaConfig)
• roberta-prelayernorm — RobertaPreLayerNormConfig (modelo RobertaPreLayerNormConfig)
• roc_bert — RoCBertConfig (modelo RoCBertConfig)
• roformer — RoFormerConfig (modelo RoFormerConfig)
• rt_detr — RTDetrConfig (modelo RTDetrConfig)
• rt_detr_resnet — RTDetrResNetConfig (modelo RTDetrResNetConfig)
• rt_detr_v2 — RTDetrV2Config (modelo RTDetrV2Config)
• rwkv — RwkvConfig (modelo RwkvConfig)
• sam — SamConfig (modelo SamConfig)
• sam2 — Sam2Config (modelo Sam2Config)
• sam2_hiera_det_model — Sam2HieraDetConfig (modelo Sam2HieraDetConfig)
• sam2_mask_decoder — Sam2MaskDecoderConfig (modelo Sam2MaskDecoderConfig)
• sam2_prompt_encoder — Sam2PromptEncoderConfig (modelo Sam2PromptEncoderConfig)
• sam2_video — Sam2VideoConfig (modelo Sam2VideoConfig)
• sam2_video_mask_decoder — Sam2VideoMaskDecoderConfig (modelo Sam2VideoMaskDecoderConfig)
• sam2_video_prompt_encoder — Sam2VideoPromptEncoderConfig (modelo Sam2VideoPromptEncoderConfig)
• sam2_vision_model — Sam2VisionConfig (modelo Sam2VisionConfig)
• sam3 — Sam3Config (modelo Sam3Config)
• sam3_detr_decoder — Sam3DETRDecoderConfig (modelo Sam3DETRDecoderConfig)
• sam3_detr_encoder — Sam3DETREncoderConfig (modelo Sam3DETREncoderConfig)
• sam3_geometry_encoder — Sam3GeometryEncoderConfig (modelo Sam3GeometryEncoderConfig)
• sam3_lite_text — Sam3LiteTextConfig (modelo Sam3LiteTextConfig)
• sam3_lite_text_detr_decoder — Sam3LiteTextDETRDecoderConfig (modelo Sam3LiteTextDETRDecoderConfig)
• sam3_lite_text_detr_encoder — Sam3LiteTextDETREncoderConfig (modelo Sam3LiteTextDETREncoderConfig)
• sam3_lite_text_geometry_encoder — Sam3LiteTextGeometryEncoderConfig (modelo Sam3LiteTextGeometryEncoderConfig)
• sam3_lite_text_mask_decoder — Sam3LiteTextMaskDecoderConfig (modelo Sam3LiteTextMaskDecoderConfig)
• sam3_lite_text_text_model — Sam3LiteTextTextConfig (modelo Sam3LiteTextTextConfig)
• sam3_mask_decoder — Sam3MaskDecoderConfig (modelo Sam3MaskDecoderConfig)
• sam3_tracker — Sam3TrackerConfig (modelo Sam3TrackerConfig)
• sam3_tracker_mask_decoder — Sam3TrackerMaskDecoderConfig (modelo Sam3TrackerMaskDecoderConfig)
• sam3_tracker_prompt_encoder — Sam3TrackerPromptEncoderConfig (modelo Sam3TrackerPromptEncoderConfig)
• sam3_tracker_video — Sam3TrackerVideoConfig (modelo Sam3TrackerVideoConfig)
• sam3_tracker_video_mask_decoder — Sam3TrackerVideoMaskDecoderConfig (modelo Sam3TrackerVideoMaskDecoderConfig)
• sam3_tracker_video_prompt_encoder — Sam3TrackerVideoPromptEncoderConfig (modelo Sam3TrackerVideoPromptEncoderConfig)
• sam3_video — Sam3VideoConfig (modelo Sam3VideoConfig)
• sam3_vision_model — Sam3VisionConfig (modelo Sam3VisionConfig)
• sam3_vit_model — Sam3ViTConfig (modelo Sam3ViTConfig)
• sam_hq — SamHQConfig (modelo SamHQConfig)
• sam_hq_mask_decoder — SamHQMaskDecoderConfig (modelo SamHQMaskDecoderConfig)
• sam_hq_prompt_encoder — SamHQPromptEncoderConfig (modelo SamHQPromptEncoderConfig)
• sam_hq_vision_model — SamHQVisionConfig (modelo SamHQVisionConfig)
• sam_mask_decoder — SamMaskDecoderConfig (modelo SamMaskDecoderConfig)
• sam_prompt_encoder — SamPromptEncoderConfig (modelo SamPromptEncoderConfig)
• sam_vision_model — SamVisionConfig (modelo SamVisionConfig)
• sapiens2 — Sapiens2Config (modelo Sapiens2Config)
• sapiens2_head — Sapiens2HeadConfig (modelo Sapiens2HeadConfig)
• seamless_m4t — SeamlessM4TConfig (modelo SeamlessM4TConfig)
• seamless_m4t_v2 — SeamlessM4Tv2Config (modelo SeamlessM4Tv2Config)
• seed_oss — SeedOssConfig (modelo SeedOssConfig)
• segformer — SegformerConfig (modelo SegformerConfig)
• seggpt — SegGptConfig (modelo SegGptConfig)
• sew — SEWConfig (modelo SEWConfig)
• sew-d — SEWDConfig (modelo SEWDConfig)
• shieldgemma2 — ShieldGemma2Config (modelo ShieldGemma2Config)
• siglip — SiglipConfig (modelo SiglipConfig)
• siglip2 — Siglip2Config (modelo Siglip2Config)
• siglip2_text_model — Siglip2TextConfig (modelo Siglip2TextConfig)
• siglip2_vision_model — Siglip2VisionConfig (modelo Siglip2VisionConfig)
• siglip_text_model — SiglipTextConfig (modelo SiglipTextConfig)
• siglip_vision_model — SiglipVisionConfig (modelo SiglipVisionConfig)
• slanet — SLANetConfig (modelo SLANetConfig)
• slanext — SLANeXtConfig (modelo SLANeXtConfig)
• slanext_vision — SLANeXtVisionConfig (modelo SLANeXtVisionConfig)
• smollm3 — SmolLM3Config (modelo SmolLM3Config)
• smolvlm — SmolVLMConfig (modelo SmolVLMConfig)
• smolvlm_vision — SmolVLMVisionConfig (modelo SmolVLMVisionConfig)
• solar_open — SolarOpenConfig (modelo SolarOpenConfig)
• speech-encoder-decoder — SpeechEncoderDecoderConfig (modelo SpeechEncoderDecoderConfig)
• speech_to_text — Speech2TextConfig (modelo Speech2TextConfig)
• speecht5 — SpeechT5Config (modelo SpeechT5Config)
• speecht5_hifigan — SpeechT5HifiGanConfig (modelo SpeechT5HifiGanConfig)
• splinter — SplinterConfig (modelo SplinterConfig)
• squeezebert — SqueezeBertConfig (modelo SqueezeBertConfig)
• stablelm — StableLmConfig (modelo StableLmConfig)
• starcoder2 — Starcoder2Config (modelo Starcoder2Config)
• step3p5 — Step3p7TextConfig (modelo Step3p7TextConfig)
• step3p5_vision — Step3p7VisionConfig (modelo Step3p7VisionConfig)
• step3p7 — Step3p7Config (modelo Step3p7Config)
• superglue — SuperGlueConfig (modelo SuperGlueConfig)
• superpoint — SuperPointConfig (modelo SuperPointConfig)
• swiftformer — SwiftFormerConfig (modelo SwiftFormerConfig)
• swin — SwinConfig (modelo SwinConfig)
• swin2sr — Swin2SRConfig (modelo Swin2SRConfig)
• swinv2 — Swinv2Config (modelo Swinv2Config)
• switch_transformers — SwitchTransformersConfig (modelo SwitchTransformersConfig)
• t5 — T5Config (modelo T5Config)
• t5_gemma_module — T5GemmaModuleConfig (modelo T5GemmaModuleConfig)
• t5gemma — T5GemmaConfig (modelo T5GemmaConfig)
• t5gemma2 — T5Gemma2Config (modelo T5Gemma2Config)
• t5gemma2_decoder — T5Gemma2DecoderConfig (modelo T5Gemma2DecoderConfig)
• t5gemma2_encoder — T5Gemma2EncoderConfig (modelo T5Gemma2EncoderConfig)
• t5gemma2_text — T5Gemma2TextConfig (modelo T5Gemma2TextConfig)
• table-transformer — TableTransformerConfig (modelo TableTransformerConfig)
• tapas — TapasConfig (modelo TapasConfig)
• textnet — TextNetConfig (modelo TextNetConfig)
• time_series_transformer — TimeSeriesTransformerConfig (modelo TimeSeriesTransformerConfig)
• timesfm — TimesFmConfig (modelo TimesFmConfig)
• timesfm2_5 — TimesFm2_5Config (modelo TimesFm2_5Config)
• timesformer — TimesformerConfig (modelo TimesformerConfig)
• timm_backbone — TimmBackboneConfig (modelo TimmBackboneConfig)
• timm_wrapper — TimmWrapperConfig (modelo TimmWrapperConfig)
• tipsv2 — Tipsv2Config (modelo Tipsv2Config)
• tipsv2_dpt — Tipsv2DptConfig (modelo Tipsv2DptConfig)
• tipsv2_text_model — Tipsv2TextConfig (modelo Tipsv2TextConfig)
• tipsv2_vision_model — Tipsv2VisionConfig (modelo Tipsv2VisionConfig)
• trocr — TrOCRConfig (modelo TrOCRConfig)
• tvp — TvpConfig (modelo TvpConfig)
• udop — UdopConfig (modelo UdopConfig)
• ultrabert — UltrabertConfig (modelo UltrabertConfig)
• ultravbert — UltraVBertConfig (modelo UltraVBertConfig)
• umt5 — UMT5Config (modelo UMT5Config)
• unispeech — UniSpeechConfig (modelo UniSpeechConfig)
• unispeech-sat — UniSpeechSatConfig (modelo UniSpeechSatConfig)
• univnet — UnivNetConfig (modelo UnivNetConfig)
• unlimited_ocr — UnlimitedOcrConfig (modelo UnlimitedOcrConfig)
• unlimited_ocr_sam_vision_model — UnlimitedOcrSamVisionConfig (modelo UnlimitedOcrSamVisionConfig)
• unlimited_ocr_text — UnlimitedOcrTextConfig (modelo UnlimitedOcrTextConfig)
• unlimited_ocr_vision — UnlimitedOcrVisionConfig (modelo UnlimitedOcrVisionConfig)
• unlimited_ocr_vision_encoder — UnlimitedOcrVisionEncoderConfig (modelo UnlimitedOcrVisionEncoderConfig)
• upernet — UperNetConfig (modelo UperNetConfig)
• uvdoc — UVDocConfig (modelo UVDocConfig)
• uvdoc_backbone — UVDocBackboneConfig (modelo UVDocBackboneConfig)
• vaultgemma — VaultGemmaConfig (modelo VaultGemmaConfig)
• vibevoice — VibeVoiceConfig (modelo VibeVoiceConfig)
• vibevoice_acoustic_tokenizer — VibeVoiceAcousticTokenizerConfig (modelo VibeVoiceAcousticTokenizerConfig)
• vibevoice_acoustic_tokenizer_decoder — VibeVoiceAcousticTokenizerDecoderConfig (modelo VibeVoiceAcousticTokenizerDecoderConfig)
• vibevoice_acoustic_tokenizer_encoder — VibeVoiceAcousticTokenizerEncoderConfig (modelo VibeVoiceAcousticTokenizerEncoderConfig)
• vibevoice_asr — VibeVoiceAsrConfig (modelo VibeVoiceAsrConfig)
• vibevoice_diffusion — VibeVoiceDiffusionHeadConfig (modelo VibeVoiceDiffusionHeadConfig)
• video_llama_3 — VideoLlama3Config (modelo VideoLlama3Config)
• video_llama_3_vision — VideoLlama3VisionConfig (modelo VideoLlama3VisionConfig)
• video_llava — VideoLlavaConfig (modelo VideoLlavaConfig)
• videomae — VideoMAEConfig (modelo VideoMAEConfig)
• videomt — VideomtConfig (modelo VideomtConfig)
• videoprism — VideoPrismConfig (modelo VideoPrismConfig)
• videoprism_text_model — VideoPrismTextConfig (modelo VideoPrismTextConfig)
• videoprism_vision_model — VideoPrismVisionConfig (modelo VideoPrismVisionConfig)
• vilt — ViltConfig (modelo ViltConfig)
• vipllava — VipLlavaConfig (modelo VipLlavaConfig)
• vision-encoder-decoder — VisionEncoderDecoderConfig (modelo VisionEncoderDecoderConfig)
• vision-text-dual-encoder — VisionTextDualEncoderConfig (modelo VisionTextDualEncoderConfig)
• visual_bert — VisualBertConfig (modelo VisualBertConfig)
• vit — ViTConfig (modelo ViTConfig)
• vit_mae — ViTMAEConfig (modelo ViTMAEConfig)
• vit_msn — ViTMSNConfig (modelo ViTMSNConfig)
• vitdet — VitDetConfig (modelo VitDetConfig)
• vitmatte — VitMatteConfig (modelo VitMatteConfig)
• vitpose — VitPoseConfig (modelo VitPoseConfig)
• vitpose_backbone — VitPoseBackboneConfig (modelo VitPoseBackboneConfig)
• vits — VitsConfig (modelo VitsConfig)
• vivit — VivitConfig (modelo VivitConfig)
• vjepa2 — VJEPA2Config (modelo VJEPA2Config)
• voxtral — VoxtralConfig (modelo VoxtralConfig)
• voxtral_encoder — VoxtralEncoderConfig (modelo VoxtralEncoderConfig)
• voxtral_realtime — VoxtralRealtimeConfig (modelo VoxtralRealtimeConfig)
• voxtral_realtime_encoder — VoxtralRealtimeEncoderConfig (modelo VoxtralRealtimeEncoderConfig)
• voxtral_realtime_text — VoxtralRealtimeTextConfig (modelo VoxtralRealtimeTextConfig)
• wav2vec2 — Wav2Vec2Config (modelo Wav2Vec2Config)
• wav2vec2-bert — Wav2Vec2BertConfig (modelo Wav2Vec2BertConfig)
• wav2vec2-conformer — Wav2Vec2ConformerConfig (modelo Wav2Vec2ConformerConfig)
• wavlm — WavLMConfig (modelo WavLMConfig)
• whisper — WhisperConfig (modelo WhisperConfig)
• xclip — XCLIPConfig (modelo XCLIPConfig)
• xclip_text_model — XCLIPTextConfig (modelo XCLIPTextConfig)
• xclip_vision_model — XCLIPVisionConfig (modelo XCLIPVisionConfig)
• xcodec — XcodecConfig (modelo XcodecConfig)
• xcodec2 — Xcodec2Config (modelo Xcodec2Config)
• xglm — XGLMConfig (modelo XGLMConfig)
• xlm — XLMConfig (modelo XLMConfig)
• xlm-roberta — XLMRobertaConfig (modelo XLMRobertaConfig)
• xlm-roberta-xl — XLMRobertaXLConfig (modelo XLMRobertaXLConfig)
• xlnet — XLNetConfig (modelo XLNetConfig)
• xlstm — xLSTMConfig (modelo xLSTMConfig)
• xmod — XmodConfig (modelo XmodConfig)
• yolos — YolosConfig (modelo YolosConfig)
• yoso — YosoConfig (modelo YosoConfig)
• youtu — YoutuConfig (modelo YoutuConfig)
• zamba — ZambaConfig (modelo ZambaConfig)
• zamba2 — Zamba2Config (modelo Zamba2Config)
• zaya — ZayaConfig (modelo ZayaConfig)
• zoedepth — ZoeDepthConfig (modelo ZoeDepthConfig)

>>> from transformers import AutoConfig

>>> # Download configuration from huggingface.co and cache.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-uncased")

>>> # Download configuration from huggingface.co (user-uploaded) and cache.
>>> config = AutoConfig.from_pretrained("dbmdz/bert-base-german-cased")

>>> # If configuration file is in a directory (e.g., was saved using *save_pretrained('./test/saved_model/')*).
>>> config = AutoConfig.from_pretrained("./test/bert_saved_model/")

>>> # Load a specific configuration file.
>>> config = AutoConfig.from_pretrained("./test/bert_saved_model/my_configuration.json")

>>> # Change some config attributes when loading a pretrained config.
>>> config = AutoConfig.from_pretrained("google-bert/bert-base-uncased", output_attentions=True, foo=False)
>>> config.output_attentions
True

>>> config, unused_kwargs = AutoConfig.from_pretrained(
...     "google-bert/bert-base-uncased", output_attentions=True, foo=False, return_unused_kwargs=True
... )
>>> config.output_attentions
True

>>> unused_kwargs
{'foo': False}

Registre uma nova configuração para esta classe.

Instancie uma das classes de tokenizador da biblioteca a partir do vocabulário de um modelo pré-treinado.

A classe de tokenizer a ser instanciada é selecionada com base na propriedade model_type do objeto de configuração (seja passada como argumento ou carregada de pretrained_model_name_or_path, quando possível), ou, quando ausente, recorrendo ao casamento de padrões em pretrained_model_name_or_path:

• EvollaModel — TokenizersBackend (modelo EvollaConfig)
• aimv2 — CLIPTokenizer (modelo Aimv2Config)
• albert — AlbertTokenizer (modelo AlbertConfig)
• align — BertTokenizer (modelo AlignConfig)
• aria — TokenizersBackend (modelo AriaConfig)
• audioflamingo3 — Qwen2Tokenizer (modelo AudioFlamingo3Config)
• aya_vision — CohereTokenizer (modelo AyaVisionConfig)
• bark — BertTokenizer (modelo BarkConfig)
• bart — RobertaTokenizer (modelo BartConfig)
• beam — TokenizersBackend (modelo BeamConfig)
• bert — BertTokenizer (modelo BertConfig)
• bert-generation — BertGenerationTokenizer (modelo BertGenerationConfig)
• big_bird — BigBirdTokenizer (modelo BigBirdConfig)
• bigbird_pegasus — TokenizersBackend (modelo BigBirdPegasusConfig)
• biogpt — BioGptTokenizer (modelo BioGptConfig)
• blenderbot — BlenderbotTokenizer (modelo BlenderbotConfig)
• blenderbot-small — BlenderbotSmallTokenizer (modelo BlenderbotSmallConfig)
• blip — BertTokenizer (modelo BlipConfig)
• blip-2 — GPT2Tokenizer (modelo Blip2Config)
• bridgetower — RobertaTokenizer (modelo BridgeTowerConfig)
• bros — BertTokenizer (modelo BrosConfig)
• camembert — CamembertTokenizer (modelo CamembertConfig)
• canary — TokenizersBackend (modelo CanaryConfig)
• canine — CanineTokenizer (modelo CanineConfig)
• chinese_clip — BertTokenizer (modelo ChineseCLIPConfig)
• clap — RobertaTokenizer (modelo ClapConfig)
• clip — CLIPTokenizer (modelo CLIPConfig)
• clipseg — CLIPTokenizer (modelo CLIPSegConfig)
• clvp — ClvpTokenizer (modelo ClvpConfig)
• codegen — GPT2Tokenizer (modelo CodeGenConfig)
• cohere — CohereTokenizer (modelo CohereConfig)
• cohere2 — CohereTokenizer (modelo Cohere2Config)
• cohere_asr — TokenizersBackend (modelo CohereAsrConfig)
• colqwen2 — Qwen2Tokenizer (modelo ColQwen2Config)
• convbert — BertTokenizer (modelo ConvBertConfig)
• cosmos3_omni — Qwen2Tokenizer (modelo Cosmos3OmniConfig)
• cpmant — CpmAntTokenizer (modelo CpmAntConfig)
• ctrl — CTRLTokenizer (modelo CTRLConfig)
• data2vec-audio — Wav2Vec2CTCTokenizer (modelo Data2VecAudioConfig)
• data2vec-text — RobertaTokenizer (modelo Data2VecTextConfig)
• dbrx — GPT2Tokenizer (modelo DbrxConfig)
• deberta — DebertaTokenizer (modelo DebertaConfig)
• deberta-v2 — DebertaV2Tokenizer (modelo DebertaV2Config)
• dia — DiaTokenizer (modelo DiaConfig)
• diffusion_gemma — GemmaTokenizer (modelo DiffusionGemmaConfig)
• distilbert — BertTokenizer (modelo DistilBertConfig)
• dpr — DPRQuestionEncoderTokenizer (modelo DPRConfig)
• electra — BertTokenizer (modelo ElectraConfig)
• emu3 — GPT2Tokenizer (modelo Emu3Config)
• ernie — BertTokenizer (modelo ErnieConfig)
• ernie4_5 — TokenizersBackend (modelo Ernie4_5Config)
• ernie4_5_moe — TokenizersBackend (modelo Ernie4_5_MoeConfig)
• esm — EsmTokenizer (modelo EsmConfig)
• esmc — EsmcTokenizer (modelo EsmcConfig)
• esmfold2 — EsmcTokenizer (modelo EsmFold2Config)
• falcon_mamba — GPTNeoXTokenizer (modelo FalconMambaConfig)
• fastspeech2_conformer — None (modelo FastSpeech2ConformerConfig)
• flaubert — FlaubertTokenizer (modelo FlaubertConfig)
• flava — BertTokenizer (modelo FlavaConfig)
• flex_olmo — TokenizersBackend (modelo FlexOlmoConfig)
• florence2 — BartTokenizer (modelo Florence2Config)
• fnet — FNetTokenizer (modelo FNetConfig)
• fsmt — FSMTTokenizer (modelo FSMTConfig)
• fun_asr_nano — Qwen2Tokenizer (modelo FunAsrNanoConfig)
• funnel — FunnelTokenizer (modelo FunnelConfig)
• fuyu — TokenizersBackend (modelo FuyuConfig)
• gemma — GemmaTokenizer (modelo GemmaConfig)
• gemma2 — GemmaTokenizer (modelo Gemma2Config)
• gemma3 — GemmaTokenizer (modelo Gemma3Config)
• gemma3_text — GemmaTokenizer (modelo Gemma3TextConfig)
• gemma3n — GemmaTokenizer (modelo Gemma3nConfig)
• gemma3n_text — GemmaTokenizer (modelo Gemma3nTextConfig)
• git — BertTokenizer (modelo GitConfig)
• glm — TokenizersBackend (modelo GlmConfig)
• glm4 — TokenizersBackend (modelo Glm4Config)
• glm4_moe — TokenizersBackend (modelo Glm4MoeConfig)
• glm4_moe_lite — TokenizersBackend (modelo Glm4MoeLiteConfig)
• glm4v — TokenizersBackend (modelo Glm4vConfig)
• glm4v_moe — TokenizersBackend (modelo Glm4vMoeConfig)
• glm_image — TokenizersBackend (modelo GlmImageConfig)
• glmasr — TokenizersBackend (modelo GlmAsrConfig)
• got_ocr2 — TokenizersBackend (modelo GotOcr2Config)
• gpt-sw3 — GPTSw3Tokenizer (modelo GPT2Config)
• gpt2 — GPT2Tokenizer (modelo GPT2Config)
• gpt_bigcode — TokenizersBackend (modelo GPTBigCodeConfig)
• gpt_neo — GPT2Tokenizer (modelo GPTNeoConfig)
• gpt_neox — GPTNeoXTokenizer (modelo GPTNeoXConfig)
• gpt_neox_japanese — GPTNeoXJapaneseTokenizer (modelo GPTNeoXJapaneseConfig)
• gptj — GPT2Tokenizer (modelo GPTJConfig)
• gradium_phonon_tts — TokenizersBackend (modelo GradiumPhononTTSConfig)
• granite — TokenizersBackend (modelo GraniteConfig)
• granite_speech5_ctc — ParakeetTokenizer (modelo GraniteSpeech5CTCConfig)
• granitemoe — TokenizersBackend (modelo GraniteMoeConfig)
• granitemoehybrid — TokenizersBackend (modelo GraniteMoeHybridConfig)
• granitemoeshared — TokenizersBackend (modelo GraniteMoeSharedConfig)
• grounding-dino — BertTokenizer (modelo GroundingDinoConfig)
• groupvit — CLIPTokenizer (modelo GroupViTConfig)
• gte — XLMRobertaTokenizer (modelo GteConfig)
• hubert — Wav2Vec2CTCTokenizer (modelo HubertConfig)
• hunyuan_vl — Qwen2Tokenizer (modelo HunYuanVLConfig)
• hyperclovax — TokenizersBackend (modelo HyperCLOVAXConfig)
• hyperclovax_vision_v2 — GPT2Tokenizer (modelo HyperCLOVAXVisionV2Config)
• ibert — RobertaTokenizer (modelo IBertConfig)
• idefics — LlamaTokenizer (modelo IdeficsConfig)
• idefics2 — LlamaTokenizer (modelo Idefics2Config)
• instructblip — GPT2Tokenizer (modelo InstructBlipConfig)
• instructblipvideo — GPT2Tokenizer (modelo InstructBlipVideoConfig)
• internvl — Qwen2Tokenizer (modelo InternVLConfig)
• jais2 — GPT2Tokenizer (modelo Jais2Config)
• jamba — TokenizersBackend (modelo JambaConfig)
• jina_embeddings_v3 — XLMRobertaTokenizer (modelo JinaEmbeddingsV3Config)
• kimi_k25 — TokenizersBackend (modelo Kimi_K25Config)
• kimi_linear — TokenizersBackend (modelo KimiLinearConfig)
• kosmos-2 — TokenizersBackend (modelo Kosmos2Config)
• lasr_ctc — LasrTokenizer (modelo LasrCTCConfig)
• lasr_encoder — LasrTokenizer (modelo LasrEncoderConfig)
• layoutlm — BertTokenizer (modelo LayoutLMConfig)
• layoutlmv2 — LayoutLMv2Tokenizer (modelo LayoutLMv2Config)
• layoutlmv3 — LayoutLMv3Tokenizer (modelo LayoutLMv3Config)
• layoutxlm — LayoutXLMTokenizer (modelo LayoutXLMConfig)
• led — LEDTokenizer (modelo LEDConfig)
• lighton_ocr — Qwen2TokenizerFast (modelo LightOnOcrConfig)
• lilt — RobertaTokenizer (modelo LiltConfig)
• llava — TokenizersBackend (modelo LlavaConfig)
• llava_next — TokenizersBackend (modelo LlavaNextConfig)
• longformer — RobertaTokenizer (modelo LongformerConfig)
• luke — LukeTokenizer (modelo LukeConfig)
• lxmert — LxmertTokenizer (modelo LxmertConfig)
• m2m_100 — M2M100Tokenizer (modelo M2M100Config)
• mamba — GPTNeoXTokenizer (modelo MambaConfig)
• mamba2 — GPTNeoXTokenizer (modelo Mamba2Config)
• marian — MarianTokenizer (modelo MarianConfig)
• markuplm — MarkupLMTokenizer (modelo MarkupLMConfig)
• mbart — MBartTokenizer (modelo MBartConfig)
• megatron-bert — BertTokenizer (modelo MegatronBertConfig)
• metaclip_2 — XLMRobertaTokenizer (modelo MetaClip2Config)
• mgp-str — MgpstrTokenizer (modelo MgpstrConfig)
• mimo_v2_flash — TokenizersBackend (modelo MiMoV2FlashConfig)
• minicpm3 — TokenizersBackend (modelo MiniCPM3Config)
• minicpm_ocr — TokenizersBackend (modelo MiniCPMOcrConfig)
• minicpmv4_6 — TokenizersBackend (modelo MiniCPMV4_6Config)
• minicpmv4_7 — TokenizersBackend (modelo MiniCPMV4_7Config)
• minimax_m2 — TokenizersBackend (modelo MiniMaxM2Config)
• ministral — MistralCommonBackend (modelo MinistralConfig)
• ministral3 — MistralCommonBackend (modelo Ministral3Config)
• mistral — MistralCommonBackend (modelo MistralConfig)
• mistral3 — MistralCommonBackend (modelo Mistral3Config)
• mistral_large4 — MistralCommonBackend (modelo MistralLarge4Config)
• mixtral — MistralCommonBackend (modelo MixtralConfig)
• mm-grounding-dino — BertTokenizer (modelo MMGroundingDinoConfig)
• mobilebert — MobileBertTokenizer (modelo MobileBertConfig)
• modernbert — TokenizersBackend (modelo ModernBertConfig)
• modernbert-decoder — TokenizersBackend (modelo ModernBertDecoderConfig)
• molmo2 — TokenizersBackend (modelo Molmo2Config)
• moss_transcribe_diarize — Qwen2Tokenizer (modelo MossTranscribeDiarizeConfig)
• mpnet — MPNetTokenizer (modelo MPNetConfig)
• mpt — GPTNeoXTokenizer (modelo MptConfig)
• mra — RobertaTokenizer (modelo MraConfig)
• mt5 — T5Tokenizer (modelo MT5Config)
• musicgen — T5Tokenizer (modelo MusicgenConfig)
• musicgen_melody — T5Tokenizer (modelo MusicgenMelodyConfig)
• mvp — MvpTokenizer (modelo MvpConfig)
• nemo_speechlm — TokenizersBackend (modelo Nemotron3_5TranscribeConfig)
• nemotron — TokenizersBackend (modelo NemotronConfig)
• nemotron3_5_asr — ParakeetTokenizer (modelo Nemotron3_5AsrConfig)
• nemotron_asr_streaming — ParakeetTokenizer (modelo NemotronAsrStreamingConfig)
• nllb-moe — NllbTokenizer (modelo NllbMoeConfig)
• nomic_bert — BertTokenizer (modelo NomicBertConfig)
• nougat — NougatTokenizer (modelo NougatConfig)
• nystromformer — TokenizersBackend (modelo NystromformerConfig)
• olmo — GPTNeoXTokenizer (modelo OlmoConfig)
• olmo2 — TokenizersBackend (modelo Olmo2Config)
• olmo3 — TokenizersBackend (modelo Olmo3Config)
• olmo_hybrid — TokenizersBackend (modelo OlmoHybridConfig)
• olmoe — GPTNeoXTokenizer (modelo OlmoeConfig)
• omdet-turbo — CLIPTokenizer (modelo OmDetTurboConfig)
• oneformer — CLIPTokenizer (modelo OneFormerConfig)
• openai-gpt — OpenAIGPTTokenizer (modelo OpenAIGPTConfig)
• opt — GPT2Tokenizer (modelo OPTConfig)
• ovis2 — Qwen2Tokenizer (modelo Ovis2Config)
• owlv2 — CLIPTokenizer (modelo Owlv2Config)
• owlvit — CLIPTokenizer (modelo OwlViTConfig)
• paddleocr_vl — TokenizersBackend (modelo PaddleOCRVLConfig)
• parakeet_ctc — ParakeetTokenizer (modelo ParakeetCTCConfig)
• parakeet_rnnt — ParakeetTokenizer (modelo ParakeetRNNTConfig)
• parakeet_tdt — ParakeetTokenizer (modelo ParakeetTDTConfig)
• pegasus — PegasusTokenizer (modelo PegasusConfig)
• pegasus_x — TokenizersBackend (modelo PegasusXConfig)
• perceiver — PerceiverTokenizer (modelo PerceiverConfig)
• persimmon — TokenizersBackend (modelo PersimmonConfig)
• phi — GPT2Tokenizer (modelo PhiConfig)
• phi3 — TokenizersBackend (modelo Phi3Config)
• phimoe — TokenizersBackend (modelo PhimoeConfig)
• pix2struct — TokenizersBackend (modelo Pix2StructConfig)
• pixtral — MistralCommonBackend (modelo PixtralVisionConfig)
• plbart — PLBartTokenizer (modelo PLBartConfig)
• pocket_tts — TokenizersBackend (modelo PocketTTSConfig)
• pp_formulanet — NougatTokenizer (modelo PPFormulaNetConfig)
• prophetnet — ProphetNetTokenizer (modelo ProphetNetConfig)
• qianfan_ocr — Qwen2Tokenizer (modelo QianfanOCRConfig)
• qwen2 — Qwen2Tokenizer (modelo Qwen2Config)
• qwen2_5_omni — Qwen2Tokenizer (modelo Qwen2_5OmniConfig)
• qwen2_5_vl — Qwen2Tokenizer (modelo Qwen2_5_VLConfig)
• qwen2_audio — Qwen2Tokenizer (modelo Qwen2AudioConfig)
• qwen2_moe — Qwen2Tokenizer (modelo Qwen2MoeConfig)
• qwen2_vl — Qwen2Tokenizer (modelo Qwen2VLConfig)
• qwen3 — Qwen2Tokenizer (modelo Qwen3Config)
• qwen3_5 — Qwen3_5Tokenizer (modelo Qwen3_5Config)
• qwen3_5_moe — Qwen3_5Tokenizer (modelo Qwen3_5MoeConfig)
• qwen3_5_moe_text — Qwen3_5Tokenizer (modelo Qwen3_5MoeTextConfig)
• qwen3_5_text — Qwen3_5Tokenizer (modelo Qwen3_5TextConfig)
• qwen3_asr — Qwen2Tokenizer (modelo Qwen3ASRConfig)
• qwen3_moe — Qwen2Tokenizer (modelo Qwen3MoeConfig)
• qwen3_next — Qwen2Tokenizer (modelo Qwen3NextConfig)
• qwen3_omni_moe — Qwen2Tokenizer (modelo Qwen3OmniMoeConfig)
• qwen3_vl — Qwen2Tokenizer (modelo Qwen3VLConfig)
• qwen3_vl_moe — Qwen2Tokenizer (modelo Qwen3VLMoeConfig)
• qwen4_exp — Qwen3_5Tokenizer (modelo Qwen4ExpConfig)
• rag — RagTokenizer (modelo RagConfig)
• recurrent_gemma — GemmaTokenizer (modelo RecurrentGemmaConfig)
• reformer — ReformerTokenizer (modelo ReformerConfig)
• rembert — RemBertTokenizer (modelo RemBertConfig)
• roberta — RobertaTokenizer (modelo RobertaConfig)
• roberta-prelayernorm — RobertaTokenizer (modelo RobertaPreLayerNormConfig)
• roc_bert — RoCBertTokenizer (modelo RoCBertConfig)
• roformer — RoFormerTokenizer (modelo RoFormerConfig)
• rwkv — GPTNeoXTokenizer (modelo RwkvConfig)
• sam3 — CLIPTokenizer (modelo Sam3Config)
• sam3_video — CLIPTokenizer (modelo Sam3VideoConfig)
• seamless_m4t — SeamlessM4TTokenizer (modelo SeamlessM4TConfig)
• seamless_m4t_v2 — SeamlessM4TTokenizer (modelo SeamlessM4Tv2Config)
• shieldgemma2 — GemmaTokenizer (modelo ShieldGemma2Config)
• siglip — SiglipTokenizer (modelo SiglipConfig)
• siglip2 — Siglip2Tokenizer (modelo Siglip2Config)
• smolvlm — TokenizersBackend (modelo SmolVLMConfig)
• speech_to_text — Speech2TextTokenizer (modelo Speech2TextConfig)
• speecht5 — SpeechT5Tokenizer (modelo SpeechT5Config)
• splinter — SplinterTokenizer (modelo SplinterConfig)
• squeezebert — BertTokenizer (modelo SqueezeBertConfig)
• stablelm — TokenizersBackend (modelo StableLmConfig)
• starcoder2 — GPT2Tokenizer (modelo Starcoder2Config)
• switch_transformers — T5Tokenizer (modelo SwitchTransformersConfig)
• t5 — T5Tokenizer (modelo T5Config)
• t5gemma — GemmaTokenizer (modelo T5GemmaConfig)
• tapas — TapasTokenizer (modelo TapasConfig)
• tipsv2 — Tipsv2Tokenizer (modelo Tipsv2Config)
• trocr — XLMRobertaTokenizer (modelo TrOCRConfig)
• tvp — BertTokenizer (modelo TvpConfig)
• udop — UdopTokenizer (modelo UdopConfig)
• umt5 — TokenizersBackend (modelo UMT5Config)
• unispeech — Wav2Vec2CTCTokenizer (modelo UniSpeechConfig)
• unispeech-sat — Wav2Vec2CTCTokenizer (modelo UniSpeechSatConfig)
• vibevoice — Qwen2TokenizerFast (modelo VibeVoiceConfig)
• videoprism — VideoPrismTokenizer (modelo VideoPrismConfig)
• vilt — BertTokenizer (modelo ViltConfig)
• vipllava — TokenizersBackend (modelo VipLlavaConfig)
• visual_bert — BertTokenizer (modelo VisualBertConfig)
• vits — VitsTokenizer (modelo VitsConfig)
• voxtral — MistralCommonBackend (modelo VoxtralConfig)
• voxtral_realtime — MistralCommonBackend (modelo VoxtralRealtimeConfig)
• wav2vec2 — Wav2Vec2CTCTokenizer (modelo Wav2Vec2Config)
• wav2vec2-bert — Wav2Vec2CTCTokenizer (modelo Wav2Vec2BertConfig)
• wav2vec2-conformer — Wav2Vec2CTCTokenizer (modelo Wav2Vec2ConformerConfig)
• whisper — WhisperTokenizer (modelo WhisperConfig)
• xclip — CLIPTokenizer (modelo XCLIPConfig)
• xglm — XGLMTokenizer (modelo XGLMConfig)
• xlm — XLMTokenizer (modelo XLMConfig)
• xlm-roberta — XLMRobertaTokenizer (modelo XLMRobertaConfig)
• xlm-roberta-xl — TokenizersBackend (modelo XLMRobertaXLConfig)
• xlnet — XLNetTokenizer (modelo XLNetConfig)
• xlstm — GPTNeoXTokenizer (modelo xLSTMConfig)
• xmod — XLMRobertaTokenizer (modelo XmodConfig)
• yoso — AlbertTokenizer (modelo YosoConfig)
• zaya — GemmaTokenizer (modelo ZayaConfig)

>>> from transformers import AutoTokenizer

>>> # Download vocabulary from huggingface.co and cache.
>>> tokenizer = AutoTokenizer.from_pretrained("google-bert/bert-base-uncased")

>>> # Download vocabulary from huggingface.co (user-uploaded) and cache.
>>> tokenizer = AutoTokenizer.from_pretrained("dbmdz/bert-base-german-cased")

>>> # If vocabulary files are in a directory (e.g. tokenizer was saved using *save_pretrained('./test/saved_model/')*)
>>> # tokenizer = AutoTokenizer.from_pretrained("./test/bert_saved_model/")

>>> # Download vocabulary from huggingface.co and define model-specific arguments
>>> tokenizer = AutoTokenizer.from_pretrained("FacebookAI/roberta-base", add_prefix_space=True)

Registre um novo tokenizer neste mapeamento.

Instancie uma das classes de extrator de recursos da biblioteca a partir do vocabulário de um modelo pré-treinado.

A classe de extrator de recursos a ser instanciada é selecionada com base na propriedade model_type do objeto de configuração (seja passada como argumento ou carregada de pretrained_model_name_or_path, quando possível), ou, quando essa propriedade não está presente, recorrendo ao casamento de padrões em pretrained_model_name_or_path:

• audio-spectrogram-transformer — ASTFeatureExtractor (modelo ASTConfig)
• audioflamingo3 — WhisperFeatureExtractor (modelo AudioFlamingo3Config)
• canary — ParakeetFeatureExtractor (modelo CanaryConfig)
• clap — ClapFeatureExtractor (modelo ClapConfig)
• clvp — ClvpFeatureExtractor (modelo ClvpConfig)
• cohere_asr — CohereAsrFeatureExtractor (modelo CohereAsrConfig)
• csm — EncodecFeatureExtractor (modelo CsmConfig)
• dac — DacFeatureExtractor (modelo DacConfig)
• data2vec-audio — Wav2Vec2FeatureExtractor (modelo Data2VecAudioConfig)
• dia — DiaFeatureExtractor (modelo DiaConfig)
• embedding_gemma2 — Gemma4AudioFeatureExtractor (modelo EmbeddingGemma2Config)
• encodec — EncodecFeatureExtractor (modelo EncodecConfig)
• fun_asr_nano — FunAsrNanoFeatureExtractor (modelo FunAsrNanoConfig)
• gemma3n — Gemma3nAudioFeatureExtractor (modelo Gemma3nConfig)
• gemma4 — Gemma4AudioFeatureExtractor (modelo Gemma4Config)
• gemma4_unified — Gemma4UnifiedAudioFeatureExtractor (modelo Gemma4UnifiedConfig)
• glmasr — WhisperFeatureExtractor (modelo GlmAsrConfig)
• gradium_phonon_tts — PocketTTSFeatureExtractor (modelo GradiumPhononTTSConfig)
• granite_speech — GraniteSpeechFeatureExtractor (modelo GraniteSpeechConfig)
• granite_speech5_ctc — GraniteSpeech5FeatureExtractor (modelo GraniteSpeech5CTCConfig)
• granite_speech5_encoder — GraniteSpeech5FeatureExtractor (modelo GraniteSpeech5EncoderConfig)
• granite_speech_plus — GraniteSpeechFeatureExtractor (modelo GraniteSpeechPlusConfig)
• higgs_audio_v2_tokenizer — DacFeatureExtractor (modelo HiggsAudioV2TokenizerConfig)
• hubert — Wav2Vec2FeatureExtractor (modelo HubertConfig)
• inkling_mm_model — InklingFeatureExtractor (modelo InklingConfig)
• kyutai_speech_to_text — KyutaiSpeechToTextFeatureExtractor (modelo KyutaiSpeechToTextConfig)
• lasr_ctc — LasrFeatureExtractor (modelo LasrCTCConfig)
• lasr_encoder — LasrFeatureExtractor (modelo LasrEncoderConfig)
• mimi — EncodecFeatureExtractor (modelo MimiConfig)
• moonshine — Wav2Vec2FeatureExtractor (modelo MoonshineConfig)
• moshi — EncodecFeatureExtractor (modelo MoshiConfig)
• moss_transcribe_diarize — MossTranscribeDiarizeFeatureExtractor (modelo MossTranscribeDiarizeConfig)
• muse_spark — MuseSparkFeatureExtractor (modelo MuseSparkConfig)
• musicgen — EncodecFeatureExtractor (modelo MusicgenConfig)
• musicgen_melody — MusicgenMelodyFeatureExtractor (modelo MusicgenMelodyConfig)
• nemo_speechlm — NemotronAsrStreamingFeatureExtractor (modelo Nemotron3_5TranscribeConfig)
• nemotron3_5_asr — NemotronAsrStreamingFeatureExtractor (modelo Nemotron3_5AsrConfig)
• nemotron3_diarization — NemotronAsrStreamingFeatureExtractor (modelo Nemotron3DiarizationConfig)
• nemotron_asr_streaming — NemotronAsrStreamingFeatureExtractor (modelo NemotronAsrStreamingConfig)
• nemotron_asr_streaming_encoder — NemotronAsrStreamingFeatureExtractor (modelo NemotronAsrStreamingEncoderConfig)
• neucodec — NeuCodecFeatureExtractor (modelo NeuCodecConfig)
• parakeet_ctc — ParakeetFeatureExtractor (modelo ParakeetCTCConfig)
• parakeet_encoder — ParakeetFeatureExtractor (modelo ParakeetEncoderConfig)
• parakeet_rnnt — ParakeetFeatureExtractor (modelo ParakeetRNNTConfig)
• parakeet_tdt — ParakeetFeatureExtractor (modelo ParakeetTDTConfig)
• pe_audio — PeAudioFeatureExtractor (modelo PeAudioConfig)
• pe_audio_video — PeAudioFeatureExtractor (modelo PeAudioVideoConfig)
• phi4_multimodal — Phi4MultimodalFeatureExtractor (modelo Phi4MultimodalConfig)
• pocket_tts — PocketTTSFeatureExtractor (modelo PocketTTSConfig)
• pocket_tts_audio_tokenizer — EncodecFeatureExtractor (modelo PocketTTSAudioTokenizerConfig)
• pop2piano — Pop2PianoFeatureExtractor (modelo Pop2PianoConfig)
• qwen2_5_omni — WhisperFeatureExtractor (modelo Qwen2_5OmniConfig)
• qwen2_audio — WhisperFeatureExtractor (modelo Qwen2AudioConfig)
• qwen3_asr — Qwen3ASRFeatureExtractor (modelo Qwen3ASRConfig)
• qwen3_omni_moe — WhisperFeatureExtractor (modelo Qwen3OmniMoeConfig)
• seamless_m4t — SeamlessM4TFeatureExtractor (modelo SeamlessM4TConfig)
• seamless_m4t_v2 — SeamlessM4TFeatureExtractor (modelo SeamlessM4Tv2Config)
• sew — Wav2Vec2FeatureExtractor (modelo SEWConfig)
• sew-d — Wav2Vec2FeatureExtractor (modelo SEWDConfig)
• speech_to_text — Speech2TextFeatureExtractor (modelo Speech2TextConfig)
• speecht5 — SpeechT5FeatureExtractor (modelo SpeechT5Config)
• unispeech — Wav2Vec2FeatureExtractor (modelo UniSpeechConfig)
• unispeech-sat — Wav2Vec2FeatureExtractor (modelo UniSpeechSatConfig)
• univnet — UnivNetFeatureExtractor (modelo UnivNetConfig)
• vibevoice — VibeVoiceAcousticTokenizerFeatureExtractor (modelo VibeVoiceConfig)
• vibevoice_acoustic_tokenizer — VibeVoiceAcousticTokenizerFeatureExtractor (modelo VibeVoiceAcousticTokenizerConfig)
• vibevoice_asr — VibeVoiceAcousticTokenizerFeatureExtractor (modelo VibeVoiceAsrConfig)
• voxtral — WhisperFeatureExtractor (modelo VoxtralConfig)
• voxtral_realtime — VoxtralRealtimeFeatureExtractor (modelo VoxtralRealtimeConfig)
• wav2vec2 — Wav2Vec2FeatureExtractor (modelo Wav2Vec2Config)
• wav2vec2-bert — Wav2Vec2FeatureExtractor (modelo Wav2Vec2BertConfig)
• wav2vec2-conformer — Wav2Vec2FeatureExtractor (modelo Wav2Vec2ConformerConfig)
• wavlm — Wav2Vec2FeatureExtractor (modelo WavLMConfig)
• whisper — WhisperFeatureExtractor (modelo WhisperConfig)
• xcodec — DacFeatureExtractor (modelo XcodecConfig)
• xcodec2 — Xcodec2FeatureExtractor (modelo Xcodec2Config)

É necessário passar token=True quando você deseja usar um modelo privado.

>>> from transformers import AutoFeatureExtractor

>>> # Download feature extractor from huggingface.co and cache.
>>> feature_extractor = AutoFeatureExtractor.from_pretrained("facebook/wav2vec2-base-960h")

>>> # If feature extractor files are in a directory (e.g. feature extractor was saved using *save_pretrained('./test/saved_model/')*)
>>> # feature_extractor = AutoFeatureExtractor.from_pretrained("./test/saved_model/")

Registre um novo extrator de recursos para esta classe.

Instancie uma das classes de processador de imagem da biblioteca a partir de um vocabulário de modelo pré-treinado.

A classe de processador de imagem a ser instanciada é selecionada com base na propriedade model_type do objeto de configuração (seja passada como argumento ou carregada de pretrained_model_name_or_path, quando possível), ou, quando essa propriedade não está presente, recorrendo ao casamento de padrões em pretrained_model_name_or_path:

• aimv2 — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo Aimv2Config)
• aimv2_vision_model — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo Aimv2VisionConfig)
• align — {'torchvision': 'EfficientNetImageProcessor', 'pil': 'EfficientNetImageProcessorPil'} (modelo AlignConfig)
• altclip — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo AltCLIPConfig)
• aria — {'pil': 'AriaImageProcessorPil', 'torchvision': 'AriaImageProcessor'} (modelo AriaConfig)
• aya_vision — {'torchvision': 'GotOcr2ImageProcessor', 'pil': 'GotOcr2ImageProcessorPil'} (modelo AyaVisionConfig)
• beit — {'pil': 'BeitImageProcessorPil', 'torchvision': 'BeitImageProcessor'} (modelo BeitConfig)
• bit — {'pil': 'BitImageProcessorPil', 'torchvision': 'BitImageProcessor'} (modelo BitConfig)
• blip — {'pil': 'BlipImageProcessorPil', 'torchvision': 'BlipImageProcessor'} (modelo BlipConfig)
• blip-2 — {'torchvision': 'BlipImageProcessor', 'pil': 'BlipImageProcessorPil'} (modelo Blip2Config)
• bridgetower — {'pil': 'BridgeTowerImageProcessorPil', 'torchvision': 'BridgeTowerImageProcessor'} (modelo BridgeTowerConfig)
• chameleon — {'pil': 'ChameleonImageProcessorPil', 'torchvision': 'ChameleonImageProcessor'} (modelo ChameleonConfig)
• chinese_clip — {'pil': 'ChineseCLIPImageProcessorPil', 'torchvision': 'ChineseCLIPImageProcessor'} (modelo ChineseCLIPConfig)
• chmv2 — {'torchvision': 'CHMv2ImageProcessor'} (modelo CHMv2Config)
• clip — {'pil': 'CLIPImageProcessorPil', 'torchvision': 'CLIPImageProcessor'} (modelo CLIPConfig)
• clipseg — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo CLIPSegConfig)
• cohere2_vision — {'torchvision': 'Cohere2VisionImageProcessor'} (modelo Cohere2VisionConfig)
• cohere_compass — {'pil': 'CohereCompassImageProcessorPil', 'torchvision': 'CohereCompassImageProcessor'} (modelo CohereCompassConfig)
• colpali — {'torchvision': 'SiglipImageProcessor', 'pil': 'SiglipImageProcessorPil'} (modelo ColPaliConfig)
• colqwen2 — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo ColQwen2Config)
• conditional_detr — {'pil': 'ConditionalDetrImageProcessorPil', 'torchvision': 'ConditionalDetrImageProcessor'} (modelo ConditionalDetrConfig)
• convnext — {'pil': 'ConvNextImageProcessorPil', 'torchvision': 'ConvNextImageProcessor'} (modelo ConvNextConfig)
• convnextv2 — {'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'} (modelo ConvNextV2Config)
• cosmos3_edge — {'pil': 'Cosmos3EdgeImageProcessorPil', 'torchvision': 'Cosmos3EdgeImageProcessor'} (modelo Cosmos3EdgeConfig)
• cosmos3_omni — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Cosmos3OmniConfig)
• cvt — {'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'} (modelo CvtConfig)
• data2vec-vision — {'torchvision': 'BeitImageProcessor', 'pil': 'BeitImageProcessorPil'} (modelo Data2VecVisionConfig)
• deepseek_ocr2 — {'pil': 'DeepseekOcr2ImageProcessorPil', 'torchvision': 'DeepseekOcr2ImageProcessor'} (modelo DeepseekOcr2Config)
• deepseek_vl — {'pil': 'DeepseekVLImageProcessorPil', 'torchvision': 'DeepseekVLImageProcessor'} (modelo DeepseekVLConfig)
• deepseek_vl_hybrid — {'pil': 'DeepseekVLHybridImageProcessorPil', 'torchvision': 'DeepseekVLHybridImageProcessor'} (modelo DeepseekVLHybridConfig)
• deformable_detr — {'pil': 'DeformableDetrImageProcessorPil', 'torchvision': 'DeformableDetrImageProcessor'} (modelo DeformableDetrConfig)
• deimv2 — {'torchvision': 'RTDetrImageProcessor', 'pil': 'RTDetrImageProcessorPil'} (modelo Deimv2Config)
• deit — {'pil': 'DeiTImageProcessorPil', 'torchvision': 'DeiTImageProcessor'} (modelo DeiTConfig)
• depth_anything — {'torchvision': 'DPTImageProcessor', 'pil': 'DPTImageProcessorPil'} (modelo DepthAnythingConfig)
• depth_pro — {'torchvision': 'DepthProImageProcessor'} (modelo DepthProConfig)
• detr — {'pil': 'DetrImageProcessorPil', 'torchvision': 'DetrImageProcessor'} (modelo DetrConfig)
• dinat — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo DinatConfig)
• dinov2 — {'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'} (modelo Dinov2Config)
• dinov3_vit — {'torchvision': 'DINOv3ViTImageProcessor'} (modelo DINOv3ViTConfig)
• donut-swin — {'torchvision': 'DonutImageProcessor', 'pil': 'DonutImageProcessorPil'} (modelo DonutSwinConfig)
• dpt — {'pil': 'DPTImageProcessorPil', 'torchvision': 'DPTImageProcessor'} (modelo DPTConfig)
• edgetam — {'torchvision': 'Sam2ImageProcessor'} (modelo EdgeTamConfig)
• efficientloftr — {'pil': 'EfficientLoFTRImageProcessorPil', 'torchvision': 'EfficientLoFTRImageProcessor'} (modelo EfficientLoFTRConfig)
• efficientnet — {'pil': 'EfficientNetImageProcessorPil', 'torchvision': 'EfficientNetImageProcessor'} (modelo EfficientNetConfig)
• embedding_gemma2 — {'pil': 'Gemma4ImageProcessorPil', 'torchvision': 'Gemma4ImageProcessor'} (modelo EmbeddingGemma2Config)
• emu3 — {'pil': 'Emu3ImageProcessor'} (modelo Emu3Config)
• eomt — {'pil': 'EomtImageProcessorPil', 'torchvision': 'EomtImageProcessor'} (modelo EomtConfig)
• eomt_dinov3 — {'torchvision': 'EomtImageProcessor', 'pil': 'EomtImageProcessorPil'} (modelo EomtDinov3Config)
• ernie4_5_vl_moe — {'pil': 'Ernie4_5_VLMoeImageProcessorPil', 'torchvision': 'Ernie4_5_VLMoeImageProcessor'} (modelo Ernie4_5_VLMoeConfig)
• exaone4_5 — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Exaone4_5_Config)
• flava — {'pil': 'FlavaImageProcessorPil', 'torchvision': 'FlavaImageProcessor'} (modelo FlavaConfig)
• florence2 — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo Florence2Config)
• focalnet — {'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'} (modelo FocalNetConfig)
• fuyu — {'pil': 'FuyuImageProcessorPil', 'torchvision': 'FuyuImageProcessor'} (modelo FuyuConfig)
• gemma3 — {'pil': 'Gemma3ImageProcessorPil', 'torchvision': 'Gemma3ImageProcessor'} (modelo Gemma3Config)
• gemma3n — {'torchvision': 'SiglipImageProcessor', 'pil': 'SiglipImageProcessorPil'} (modelo Gemma3nConfig)
• gemma4 — {'pil': 'Gemma4ImageProcessorPil', 'torchvision': 'Gemma4ImageProcessor'} (modelo Gemma4Config)
• gemma4_unified — {'torchvision': 'Gemma4UnifiedImageProcessor'} (modelo Gemma4UnifiedConfig)
• git — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo GitConfig)
• glm46v — {'pil': 'Glm46VImageProcessorPil', 'torchvision': 'Glm46VImageProcessor'} (modelo Glm46VConfig)
• glm4v — {'pil': 'Glm4vImageProcessorPil', 'torchvision': 'Glm4vImageProcessor'} (modelo Glm4vConfig)
• glm5_next — {'pil': 'Glm5NextImageProcessorPil', 'torchvision': 'Glm5NextImageProcessor'} (modelo Glm5NextConfig)
• glm_image — {'pil': 'GlmImageImageProcessorPil', 'torchvision': 'GlmImageImageProcessor'} (modelo GlmImageConfig)
• glmga — {'pil': 'GlmgaImageProcessorPil', 'torchvision': 'GlmgaImageProcessor'} (modelo GlmgaConfig)
• glpn — {'pil': 'GLPNImageProcessorPil', 'torchvision': 'GLPNImageProcessor'} (modelo GLPNConfig)
• got_ocr2 — {'pil': 'GotOcr2ImageProcessorPil', 'torchvision': 'GotOcr2ImageProcessor'} (modelo GotOcr2Config)
• granite4_vision — {'torchvision': 'LlavaNextImageProcessor', 'pil': 'LlavaNextImageProcessorPil'} (modelo Granite4VisionConfig)
• grounding-dino — {'pil': 'GroundingDinoImageProcessorPil', 'torchvision': 'GroundingDinoImageProcessor'} (modelo GroundingDinoConfig)
• groupvit — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo GroupViTConfig)
• hiera — {'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'} (modelo HieraConfig)
• hunyuan_vl — {'pil': 'HunYuanVLImageProcessorPil', 'torchvision': 'HunYuanVLImageProcessor'} (modelo HunYuanVLConfig)
• hyperclovax_vision_v2 — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo HyperCLOVAXVisionV2Config)
• idefics — {'pil': 'IdeficsImageProcessorPil', 'torchvision': 'IdeficsImageProcessor'} (modelo IdeficsConfig)
• idefics2 — {'pil': 'Idefics2ImageProcessorPil', 'torchvision': 'Idefics2ImageProcessor'} (modelo Idefics2Config)
• idefics3 — {'pil': 'Idefics3ImageProcessorPil', 'torchvision': 'Idefics3ImageProcessor'} (modelo Idefics3Config)
• ijepa — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo IJepaConfig)
• imagegpt — {'pil': 'ImageGPTImageProcessorPil', 'torchvision': 'ImageGPTImageProcessor'} (modelo ImageGPTConfig)
• inkling_mm_model — {'torchvision': 'InklingImageProcessor'} (modelo InklingConfig)
• instructblip — {'torchvision': 'BlipImageProcessor', 'pil': 'BlipImageProcessorPil'} (modelo InstructBlipConfig)
• internvl — {'torchvision': 'GotOcr2ImageProcessor', 'pil': 'GotOcr2ImageProcessorPil'} (modelo InternVLConfig)
• janus — {'pil': 'JanusImageProcessorPil', 'torchvision': 'JanusImageProcessor'} (modelo JanusConfig)
• kimi_k25 — {'torchvision': 'Kimi_K25ImageProcessor'} (modelo Kimi_K25Config)
• kosmos-2 — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo Kosmos2Config)
• kosmos-2.5 — {'torchvision': 'Kosmos2_5ImageProcessor', 'pil': 'Kosmos2_5ImageProcessorPil'} (modelo Kosmos2_5Config)
• layoutlmv2 — {'pil': 'LayoutLMv2ImageProcessorPil', 'torchvision': 'LayoutLMv2ImageProcessor'} (modelo LayoutLMv2Config)
• layoutlmv3 — {'pil': 'LayoutLMv3ImageProcessorPil', 'torchvision': 'LayoutLMv3ImageProcessor'} (modelo LayoutLMv3Config)
• layoutxlm — {'torchvision': 'LayoutLMv2ImageProcessor', 'pil': 'LayoutLMv2ImageProcessorPil'} (modelo LayoutXLMConfig)
• levit — {'pil': 'LevitImageProcessorPil', 'torchvision': 'LevitImageProcessor'} (modelo LevitConfig)
• lfm2_vl — {'torchvision': 'Lfm2VlImageProcessor'} (modelo Lfm2VlConfig)
• lightglue — {'pil': 'LightGlueImageProcessorPil', 'torchvision': 'LightGlueImageProcessor'} (modelo LightGlueConfig)
• lighton_ocr — {'torchvision': 'PixtralImageProcessor', 'pil': 'PixtralImageProcessorPil'} (modelo LightOnOcrConfig)
• llama4 — {'torchvision': 'Llama4ImageProcessor'} (modelo Llama4Config)
• llava — {'pil': 'LlavaImageProcessorPil', 'torchvision': 'LlavaImageProcessor'} (modelo LlavaConfig)
• llava_next — {'pil': 'LlavaNextImageProcessorPil', 'torchvision': 'LlavaNextImageProcessor'} (modelo LlavaNextConfig)
• llava_next_video — {'torchvision': 'LlavaNextImageProcessor', 'pil': 'LlavaNextImageProcessorPil'} (modelo LlavaNextVideoConfig)
• llava_onevision — {'pil': 'LlavaOnevisionImageProcessorPil', 'torchvision': 'LlavaOnevisionImageProcessor'} (modelo LlavaOnevisionConfig)
• lw_detr — {'torchvision': 'DeformableDetrImageProcessor', 'pil': 'DeformableDetrImageProcessorPil'} (modelo LwDetrConfig)
• mask2former — {'pil': 'Mask2FormerImageProcessorPil', 'torchvision': 'Mask2FormerImageProcessor'} (modelo Mask2FormerConfig)
• maskformer — {'pil': 'MaskFormerImageProcessorPil', 'torchvision': 'MaskFormerImageProcessor'} (modelo MaskFormerConfig)
• metaclip_2 — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo MetaClip2Config)
• mgp-str — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo MgpstrConfig)
• minicpm_ocr — {'torchvision': 'MiniCPMOcrImageProcessor'} (modelo MiniCPMOcrConfig)
• minicpmv4_6 — {'pil': 'MiniCPMV4_6ImageProcessorPil', 'torchvision': 'MiniCPMV4_6ImageProcessor'} (modelo MiniCPMV4_6Config)
• minicpmv4_7 — {'pil': 'MiniCPMV4_6ImageProcessorPil', 'torchvision': 'MiniCPMV4_6ImageProcessor'} (modelo MiniCPMV4_7Config)
• minimax_m3_vl — {'torchvision': 'MiniMaxM3VLImageProcessor'} (modelo MiniMaxM3VLConfig)
• mistral3 — {'torchvision': 'PixtralImageProcessor', 'pil': 'PixtralImageProcessorPil'} (modelo Mistral3Config)
• mistral_large4 — {'torchvision': 'PixtralImageProcessor', 'pil': 'PixtralImageProcessorPil'} (modelo MistralLarge4Config)
• mlcd — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo MLCDVisionConfig)
• mllama — {'pil': 'MllamaImageProcessorPil', 'torchvision': 'MllamaImageProcessor'} (modelo MllamaConfig)
• mm-grounding-dino — {'torchvision': 'GroundingDinoImageProcessor', 'pil': 'GroundingDinoImageProcessorPil'} (modelo MMGroundingDinoConfig)
• mobilenet_v1 — {'pil': 'MobileNetV1ImageProcessorPil', 'torchvision': 'MobileNetV1ImageProcessor'} (modelo MobileNetV1Config)
• mobilenet_v2 — {'pil': 'MobileNetV2ImageProcessorPil', 'torchvision': 'MobileNetV2ImageProcessor'} (modelo MobileNetV2Config)
• mobilevit — {'pil': 'MobileViTImageProcessorPil', 'torchvision': 'MobileViTImageProcessor'} (modelo MobileViTConfig)
• mobilevitv2 — {'torchvision': 'MobileViTImageProcessor', 'pil': 'MobileViTImageProcessorPil'} (modelo MobileViTV2Config)
• molmo2 — {'torchvision': 'Molmo2ImageProcessor'} (modelo Molmo2Config)
• muse_glimmer — {'torchvision': 'MuseGlimmerImageProcessor'} (modelo MuseGlimmerConfig)
• muse_spark — {'torchvision': 'MuseGlimmerImageProcessor'} (modelo MuseSparkConfig)
• nemotron_h_omni — {'torchvision': 'NemotronH_Omni_Reasoning_V3ImageProcessor'} (modelo NemotronH_Omni_Reasoning_V3_Config)
• neomme — {'torchvision': 'NeoMMEImageProcessor'} (modelo NeoMMEConfig)
• nougat — {'pil': 'NougatImageProcessorPil', 'torchvision': 'NougatImageProcessor'} (modelo NougatConfig)
• omdet-turbo — {'torchvision': 'DetrImageProcessor', 'pil': 'DetrImageProcessorPil'} (modelo OmDetTurboConfig)
• oneformer — {'pil': 'OneFormerImageProcessorPil', 'torchvision': 'OneFormerImageProcessor'} (modelo OneFormerConfig)
• ovis2 — {'pil': 'Ovis2ImageProcessorPil', 'torchvision': 'Ovis2ImageProcessor'} (modelo Ovis2Config)
• owlv2 — {'pil': 'Owlv2ImageProcessorPil', 'torchvision': 'Owlv2ImageProcessor'} (modelo Owlv2Config)
• owlvit — {'pil': 'OwlViTImageProcessorPil', 'torchvision': 'OwlViTImageProcessor'} (modelo OwlViTConfig)
• paddleocr_vl — {'pil': 'PaddleOCRVLImageProcessorPil', 'torchvision': 'PaddleOCRVLImageProcessor'} (modelo PaddleOCRVLConfig)
• paligemma — {'torchvision': 'SiglipImageProcessor', 'pil': 'SiglipImageProcessorPil'} (modelo PaliGemmaConfig)
• perceiver — {'pil': 'PerceiverImageProcessorPil', 'torchvision': 'PerceiverImageProcessor'} (modelo PerceiverConfig)
• perception_lm — {'torchvision': 'PerceptionLMImageProcessor'} (modelo PerceptionLMConfig)
• phi4_multimodal — {'torchvision': 'Phi4MultimodalImageProcessor'} (modelo Phi4MultimodalConfig)
• pi0 — {'torchvision': 'PI0ImageProcessor'} (modelo PI0Config)
• pix2struct — {'pil': 'Pix2StructImageProcessorPil', 'torchvision': 'Pix2StructImageProcessor'} (modelo Pix2StructConfig)
• pixio — {'torchvision': 'BitImageProcessor', 'pil': 'BitImageProcessorPil'} (modelo PixioConfig)
• pixtral — {'pil': 'PixtralImageProcessorPil', 'torchvision': 'PixtralImageProcessor'} (modelo PixtralVisionConfig)
• poolformer — {'pil': 'PoolFormerImageProcessorPil', 'torchvision': 'PoolFormerImageProcessor'} (modelo PoolFormerConfig)
• pp_chart2table — {'pil': 'PPChart2TableImageProcessorPil', 'torchvision': 'PPChart2TableImageProcessor'} (modelo PPChart2TableConfig)
• pp_doclayout_v2 — {'torchvision': 'PPDocLayoutV2ImageProcessor'} (modelo PPDocLayoutV2Config)
• pp_doclayout_v3 — {'torchvision': 'PPDocLayoutV3ImageProcessor'} (modelo PPDocLayoutV3Config)
• pp_doclayout_v4 — {'torchvision': 'PPDocLayoutV4ImageProcessor'} (modelo PPDocLayoutV4Config)
• pp_formulanet — {'torchvision': 'PPFormulaNetImageProcessor'} (modelo PPFormulaNetConfig)
• pp_lcnet — {'torchvision': 'PPLCNetImageProcessor'} (modelo PPLCNetConfig)
• pp_ocrv5_mobile_det — {'torchvision': 'PPOCRV5ServerDetImageProcessor'} (modelo PPOCRV5MobileDetConfig)
• pp_ocrv5_mobile_rec — {'torchvision': 'PPOCRV5ServerRecImageProcessor'} (modelo PPOCRV5MobileRecConfig)
• pp_ocrv5_server_det — {'torchvision': 'PPOCRV5ServerDetImageProcessor'} (modelo PPOCRV5ServerDetConfig)
• pp_ocrv5_server_rec — {'torchvision': 'PPOCRV5ServerRecImageProcessor'} (modelo PPOCRV5ServerRecConfig)
• pp_ocrv6_medium_det — {'torchvision': 'PPOCRV5ServerDetImageProcessor'} (modelo PPOCRV6MediumDetConfig)
• pp_ocrv6_small_det — {'torchvision': 'PPOCRV5ServerDetImageProcessor'} (modelo PPOCRV6SmallDetConfig)
• pp_ocrv6_small_rec — {'torchvision': 'PPOCRV6SmallRecImageProcessor'} (modelo PPOCRV6SmallRecConfig)
• pp_ocrv6_tiny_rec — {'torchvision': 'PPOCRV6SmallRecImageProcessor'} (modelo PPOCRV6TinyRecConfig)
• prompt_depth_anything — {'pil': 'PromptDepthAnythingImageProcessorPil', 'torchvision': 'PromptDepthAnythingImageProcessor'} (modelo PromptDepthAnythingConfig)
• pvt — {'pil': 'PvtImageProcessorPil', 'torchvision': 'PvtImageProcessor'} (modelo PvtConfig)
• pvt_v2 — {'torchvision': 'PvtImageProcessor', 'pil': 'PvtImageProcessorPil'} (modelo PvtV2Config)
• qianfan_ocr — {'torchvision': 'GotOcr2ImageProcessor', 'pil': 'GotOcr2ImageProcessorPil'} (modelo QianfanOCRConfig)
• qwen2_5_omni — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen2_5OmniConfig)
• qwen2_5_vl — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen2_5_VLConfig)
• qwen2_vl — {'pil': 'Qwen2VLImageProcessorPil', 'torchvision': 'Qwen2VLImageProcessor'} (modelo Qwen2VLConfig)
• qwen3_5 — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen3_5Config)
• qwen3_5_moe — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen3_5MoeConfig)
• qwen3_omni_moe — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen3OmniMoeConfig)
• qwen3_vl — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen3VLConfig)
• qwen4_exp — {'torchvision': 'Qwen2VLImageProcessor', 'pil': 'Qwen2VLImageProcessorPil'} (modelo Qwen4ExpConfig)
• regnet — {'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'} (modelo RegNetConfig)
• resnet — {'torchvision': 'ConvNextImageProcessor', 'pil': 'ConvNextImageProcessorPil'} (modelo ResNetConfig)
• rf_detr — {'torchvision': 'RfDetrImageProcessor'} (modelo RfDetrConfig)
• rt_detr — {'pil': 'RTDetrImageProcessorPil', 'torchvision': 'RTDetrImageProcessor'} (modelo RTDetrConfig)
• sam — {'pil': 'SamImageProcessorPil', 'torchvision': 'SamImageProcessor'} (modelo SamConfig)
• sam2 — {'torchvision': 'Sam2ImageProcessor'} (modelo Sam2Config)
• sam2_video — {'torchvision': 'Sam2ImageProcessor'} (modelo Sam2VideoConfig)
• sam3 — {'torchvision': 'Sam3ImageProcessor'} (modelo Sam3Config)
• sam3_lite_text — {'torchvision': 'Sam3ImageProcessor'} (modelo Sam3LiteTextConfig)
• sam3_tracker — {'torchvision': 'Sam3ImageProcessor'} (modelo Sam3TrackerConfig)
• sam3_tracker_video — {'torchvision': 'Sam3ImageProcessor'} (modelo Sam3TrackerVideoConfig)
• sam3_video — {'torchvision': 'Sam3ImageProcessor'} (modelo Sam3VideoConfig)
• sam_hq — {'torchvision': 'SamImageProcessor', 'pil': 'SamImageProcessorPil'} (modelo SamHQConfig)
• sapiens2 — {'torchvision': 'Sapiens2ImageProcessor'} (modelo Sapiens2Config)
• segformer — {'pil': 'SegformerImageProcessorPil', 'torchvision': 'SegformerImageProcessor'} (modelo SegformerConfig)
• seggpt — {'pil': 'SegGptImageProcessorPil', 'torchvision': 'SegGptImageProcessor'} (modelo SegGptConfig)
• shieldgemma2 — {'torchvision': 'Gemma3ImageProcessor', 'pil': 'Gemma3ImageProcessorPil'} (modelo ShieldGemma2Config)
• siglip — {'pil': 'SiglipImageProcessorPil', 'torchvision': 'SiglipImageProcessor'} (modelo SiglipConfig)
• siglip2 — {'pil': 'Siglip2ImageProcessorPil', 'torchvision': 'Siglip2ImageProcessor'} (modelo Siglip2Config)
• slanet — {'torchvision': 'SLANeXtImageProcessor'} (modelo SLANetConfig)
• slanext — {'torchvision': 'SLANeXtImageProcessor'} (modelo SLANeXtConfig)
• smolvlm — {'pil': 'SmolVLMImageProcessorPil', 'torchvision': 'SmolVLMImageProcessor'} (modelo SmolVLMConfig)
• step3p7 — {'torchvision': 'Step3p7ImageProcessor'} (modelo Step3p7Config)
• superglue — {'pil': 'SuperGlueImageProcessorPil', 'torchvision': 'SuperGlueImageProcessor'} (modelo SuperGlueConfig)
• superpoint — {'pil': 'SuperPointImageProcessorPil', 'torchvision': 'SuperPointImageProcessor'} (modelo SuperPointConfig)
• swiftformer — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo SwiftFormerConfig)
• swin — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo SwinConfig)
• swin2sr — {'pil': 'Swin2SRImageProcessorPil', 'torchvision': 'Swin2SRImageProcessor'} (modelo Swin2SRConfig)
• swinv2 — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo Swinv2Config)
• t5gemma2 — {'torchvision': 'Gemma3ImageProcessor', 'pil': 'Gemma3ImageProcessorPil'} (modelo T5Gemma2Config)
• t5gemma2_encoder — {'torchvision': 'Gemma3ImageProcessor', 'pil': 'Gemma3ImageProcessorPil'} (modelo T5Gemma2EncoderConfig)
• table-transformer — {'torchvision': 'DetrImageProcessor', 'pil': 'DetrImageProcessorPil'} (modelo TableTransformerConfig)
• textnet — {'pil': 'TextNetImageProcessorPil', 'torchvision': 'TextNetImageProcessor'} (modelo TextNetConfig)
• timesformer — {'pil': 'VideoMAEImageProcessorPil', 'torchvision': 'VideoMAEImageProcessor'} (modelo TimesformerConfig)
• timm_wrapper — {'pil': 'TimmWrapperImageProcessor'} (modelo TimmWrapperConfig)
• tipsv2 — {'torchvision': 'Tipsv2ImageProcessor'} (modelo Tipsv2Config)
• tipsv2_dpt — {'torchvision': 'Tipsv2DptImageProcessor'} (modelo Tipsv2DptConfig)
• trocr — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo TrOCRConfig)
• tvp — {'pil': 'TvpImageProcessorPil', 'torchvision': 'TvpImageProcessor'} (modelo TvpConfig)
• udop — {'torchvision': 'LayoutLMv3ImageProcessor', 'pil': 'LayoutLMv3ImageProcessorPil'} (modelo UdopConfig)
• ultravbert — {'torchvision': 'UltraVBertImageProcessor'} (modelo UltraVBertConfig)
• unlimited_ocr — {'torchvision': 'UnlimitedOcrImageProcessor'} (modelo UnlimitedOcrConfig)
• upernet — {'torchvision': 'SegformerImageProcessor', 'pil': 'SegformerImageProcessorPil'} (modelo UperNetConfig)
• uvdoc — {'torchvision': 'UVDocImageProcessor'} (modelo UVDocConfig)
• video_llama_3 — {'pil': 'VideoLlama3ImageProcessorPil', 'torchvision': 'VideoLlama3ImageProcessor'} (modelo VideoLlama3Config)
• video_llava — {'pil': 'VideoLlavaImageProcessor'} (modelo VideoLlavaConfig)
• videomae — {'pil': 'VideoMAEImageProcessorPil', 'torchvision': 'VideoMAEImageProcessor'} (modelo VideoMAEConfig)
• vilt — {'pil': 'ViltImageProcessorPil', 'torchvision': 'ViltImageProcessor'} (modelo ViltConfig)
• vipllava — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo VipLlavaConfig)
• vit — {'pil': 'ViTImageProcessorPil', 'torchvision': 'ViTImageProcessor'} (modelo ViTConfig)
• vit_mae — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo ViTMAEConfig)
• vit_msn — {'torchvision': 'ViTImageProcessor', 'pil': 'ViTImageProcessorPil'} (modelo ViTMSNConfig)
• vitmatte — {'pil': 'VitMatteImageProcessorPil', 'torchvision': 'VitMatteImageProcessor'} (modelo VitMatteConfig)
• vitpose — {'pil': 'VitPoseImageProcessorPil', 'torchvision': 'VitPoseImageProcessor'} (modelo VitPoseConfig)
• vivit — {'torchvision': 'VivitImageProcessor'} (modelo VivitConfig)
• xclip — {'torchvision': 'CLIPImageProcessor', 'pil': 'CLIPImageProcessorPil'} (modelo XCLIPConfig)
• yolos — {'pil': 'YolosImageProcessorPil', 'torchvision': 'YolosImageProcessor'} (modelo YolosConfig)
• zoedepth — {'pil': 'ZoeDepthImageProcessorPil', 'torchvision': 'ZoeDepthImageProcessor'} (modelo ZoeDepthConfig)

É necessário passar token=True quando você deseja usar um modelo privado.

>>> from transformers import AutoImageProcessor

>>> # Download image processor from huggingface.co and cache.
>>> image_processor = AutoImageProcessor.from_pretrained("google/vit-base-patch16-224-in21k")

>>> # If image processor files are in a directory (e.g. image processor was saved using *save_pretrained('./test/saved_model/')*)
>>> # image_processor = AutoImageProcessor.from_pretrained("./test/saved_model/")

Registre um novo processador de imagem para esta classe.

Instancie uma das classes de processador de vídeo da biblioteca a partir de um vocabulário de modelo pré-treinado.

A classe de processador de vídeo a ser instanciada é selecionada com base na propriedade model_type do objeto de configuração (seja passada como argumento ou carregada de pretrained_model_name_or_path, quando possível), ou, quando essa propriedade não está presente, recorrendo ao casamento de padrões em pretrained_model_name_or_path:

• cohere_compass — {'torchvision': 'CohereCompassVideoProcessor'} (modelo CohereCompassConfig)
• cosmos3_edge — {'torchvision': 'Cosmos3EdgeVideoProcessor'} (modelo Cosmos3EdgeConfig)
• cosmos3_omni — {'torchvision': 'Qwen3VLVideoProcessor'} (modelo Cosmos3OmniConfig)
• embedding_gemma2 — {'torchvision': 'EmbeddingGemma2VideoProcessor'} (modelo EmbeddingGemma2Config)
• ernie4_5_vl_moe — {'torchvision': 'Ernie4_5_VLMoeVideoProcessor'} (modelo Ernie4_5_VLMoeConfig)
• exaone4_5 — {'torchvision': 'Qwen2VLVideoProcessor'} (modelo Exaone4_5_Config)
• gemma4 — {'torchvision': 'Gemma4VideoProcessor'} (modelo Gemma4Config)
• gemma4_unified — {'torchvision': 'Gemma4UnifiedVideoProcessor'} (modelo Gemma4UnifiedConfig)
• glm46v — {'torchvision': 'Glm46VVideoProcessor'} (modelo Glm46VConfig)
• glm4v — {'torchvision': 'Glm4vVideoProcessor'} (modelo Glm4vConfig)
• glm5_next — {'torchvision': 'Glm5NextVideoProcessor'} (modelo Glm5NextConfig)
• glmga — {'torchvision': 'GlmgaVideoProcessor'} (modelo GlmgaConfig)
• hyperclovax_vision_v2 — {'torchvision': 'Qwen2VLVideoProcessor'} (modelo HyperCLOVAXVisionV2Config)
• instructblip — {'torchvision': 'InstructBlipVideoVideoProcessor'} (modelo InstructBlipConfig)
• instructblipvideo — {'torchvision': 'InstructBlipVideoVideoProcessor'} (modelo InstructBlipVideoConfig)
• internvl — {'torchvision': 'InternVLVideoProcessor'} (modelo InternVLConfig)
• kimi_k25 — {'torchvision': 'Kimi_K25VideoProcessor'} (modelo Kimi_K25Config)
• llava_next_video — {'torchvision': 'LlavaNextVideoVideoProcessor'} (modelo LlavaNextVideoConfig)
• llava_onevision — {'torchvision': 'LlavaOnevisionVideoProcessor'} (modelo LlavaOnevisionConfig)
• minicpm_ocr — {'torchvision': 'MiniCPMOcrVideoProcessor'} (modelo MiniCPMOcrConfig)
• minicpmv4_6 — {'torchvision': 'MiniCPMV4_6VideoProcessor'} (modelo MiniCPMV4_6Config)
• minicpmv4_7 — {'torchvision': 'MiniCPMV4_6VideoProcessor'} (modelo MiniCPMV4_7Config)
• minimax_m3_vl — {'torchvision': 'MiniMaxM3VLVideoProcessor'} (modelo MiniMaxM3VLConfig)
• molmo2 — {'torchvision': 'Molmo2VideoProcessor'} (modelo Molmo2Config)
• muse_glimmer — {'torchvision': 'MuseGlimmerVideoProcessor'} (modelo MuseGlimmerConfig)
• muse_spark — {'torchvision': 'MuseGlimmerVideoProcessor'} (modelo MuseSparkConfig)
• nemotron_h_omni — {'torchvision': 'NemotronH_Omni_Reasoning_V3VideoProcessor'} (modelo NemotronH_Omni_Reasoning_V3_Config)
• pe_audio_video — {'torchvision': 'PeVideoVideoProcessor'} (modelo PeAudioVideoConfig)
• pe_video — {'torchvision': 'PeVideoVideoProcessor'} (modelo PeVideoConfig)
• perception_lm — {'torchvision': 'PerceptionLMVideoProcessor'} (modelo PerceptionLMConfig)
• qwen2_5_omni — {'torchvision': 'Qwen2VLVideoProcessor'} (modelo Qwen2_5OmniConfig)
• qwen2_5_vl — {'torchvision': 'Qwen2VLVideoProcessor'} (modelo Qwen2_5_VLConfig)
• qwen2_vl — {'torchvision': 'Qwen2VLVideoProcessor'} (modelo Qwen2VLConfig)
• qwen3_5 — {'torchvision': 'Qwen3VLVideoProcessor'} (modelo Qwen3_5Config)
• qwen3_5_moe — {'torchvision': 'Qwen3VLVideoProcessor'} (modelo Qwen3_5MoeConfig)
• qwen3_omni_moe — {'torchvision': 'Qwen2VLVideoProcessor'} (modelo Qwen3OmniMoeConfig)
• qwen3_vl — {'torchvision': 'Qwen3VLVideoProcessor'} (modelo Qwen3VLConfig)
• qwen3_vl_moe — {'torchvision': 'Qwen3VLVideoProcessor'} (modelo Qwen3VLMoeConfig)
• qwen4_exp — {'torchvision': 'Qwen3VLVideoProcessor'} (modelo Qwen4ExpConfig)
• sam2_video — {'torchvision': 'Sam2VideoVideoProcessor'} (modelo Sam2VideoConfig)
• smolvlm — {'torchvision': 'SmolVLMVideoProcessor'} (modelo SmolVLMConfig)
• timesformer — {'torchvision': 'TimesformerVideoProcessor'} (modelo TimesformerConfig)
• video_llama_3 — {'torchvision': 'VideoLlama3VideoProcessor'} (modelo VideoLlama3Config)
• video_llava — {'torchvision': 'VideoLlavaVideoProcessor'} (modelo VideoLlavaConfig)
• videomae — {'torchvision': 'VideoMAEVideoProcessor'} (modelo VideoMAEConfig)
• videomt — {'torchvision': 'VideomtVideoProcessor'} (modelo VideomtConfig)
• videoprism — {'torchvision': 'LlavaOnevisionVideoProcessor'} (modelo VideoPrismConfig)
• vivit — {'torchvision': 'VivitVideoProcessor'} (modelo VivitConfig)
• vjepa2 — {'torchvision': 'VJEPA2VideoProcessor'} (modelo VJEPA2Config)

É necessário passar token=True quando você deseja usar um modelo privado.

>>> from transformers import AutoVideoProcessor

>>> # Download video processor from huggingface.co and cache.
>>> video_processor = AutoVideoProcessor.from_pretrained("llava-hf/llava-onevision-qwen2-0.5b-ov-hf")

>>> # If video processor files are in a directory (e.g. video processor was saved using *save_pretrained('./test/saved_model/')*)
>>> # video_processor = AutoVideoProcessor.from_pretrained("./test/saved_model/")

Registre um novo processador de vídeo para esta classe.

Instancie uma das classes de processador da biblioteca a partir do vocabulário de um modelo pré-treinado.

A classe de processador a ser instanciada é selecionada com base na propriedade model_type do objeto de configuração (seja passada como argumento ou carregada de pretrained_model_name_or_path, quando possível):

• aimv2 — CLIPProcessor (modelo Aimv2Config)
• align — AlignProcessor (modelo AlignConfig)
• altclip — AltCLIPProcessor (modelo AltCLIPConfig)
• aria — AriaProcessor (modelo AriaConfig)
• audioflamingo3 — AudioFlamingo3Processor (modelo AudioFlamingo3Config)
• aya_vision — AyaVisionProcessor (modelo AyaVisionConfig)
• bark — BarkProcessor (modelo BarkConfig)
• blip — BlipProcessor (modelo BlipConfig)
• blip-2 — Blip2Processor (modelo Blip2Config)
• bridgetower — BridgeTowerProcessor (modelo BridgeTowerConfig)
• bros — BrosProcessor (modelo BrosConfig)
• canary — CanaryProcessor (modelo CanaryConfig)
• chameleon — ChameleonProcessor (modelo ChameleonConfig)
• chinese_clip — ChineseCLIPProcessor (modelo ChineseCLIPConfig)
• clap — ClapProcessor (modelo ClapConfig)
• clip — CLIPProcessor (modelo CLIPConfig)
• clipseg — CLIPSegProcessor (modelo CLIPSegConfig)
• clvp — ClvpProcessor (modelo ClvpConfig)
• cohere2_vision — Cohere2VisionProcessor (modelo Cohere2VisionConfig)
• cohere_asr — CohereAsrProcessor (modelo CohereAsrConfig)
• cohere_compass — CohereCompassProcessor (modelo CohereCompassConfig)
• colmodernvbert — ColModernVBertProcessor (modelo ColModernVBertConfig)
• colpali — ColPaliProcessor (modelo ColPaliConfig)
• colqwen2 — ColQwen2Processor (modelo ColQwen2Config)
• cosmos3_edge — Cosmos3EdgeProcessor (modelo Cosmos3EdgeConfig)
• csm — CsmProcessor (modelo CsmConfig)
• deepseek_ocr2 — DeepseekOcr2Processor (modelo DeepseekOcr2Config)
• deepseek_vl — DeepseekVLProcessor (modelo DeepseekVLConfig)
• deepseek_vl_hybrid — DeepseekVLHybridProcessor (modelo DeepseekVLHybridConfig)
• dia — DiaProcessor (modelo DiaConfig)
• diffusion_gemma — Gemma4Processor (modelo DiffusionGemmaConfig)
• edgetam — Sam2Processor (modelo EdgeTamConfig)
• embedding_gemma2 — EmbeddingGemma2Processor (modelo EmbeddingGemma2Config)
• emu3 — Emu3Processor (modelo Emu3Config)
• ernie4_5_vl_moe — Ernie4_5_VLMoeProcessor (modelo Ernie4_5_VLMoeConfig)
• evolla — EvollaProcessor (modelo EvollaConfig)
• exaone4_5 — Exaone4_5_Processor (modelo Exaone4_5_Config)
• flava — FlavaProcessor (modelo FlavaConfig)
• florence2 — Florence2Processor (modelo Florence2Config)
• fun_asr_nano — FunAsrNanoProcessor (modelo FunAsrNanoConfig)
• fuyu — FuyuProcessor (modelo FuyuConfig)
• gemma3 — Gemma3Processor (modelo Gemma3Config)
• gemma3n — Gemma3nProcessor (modelo Gemma3nConfig)
• gemma4 — Gemma4Processor (modelo Gemma4Config)
• gemma4_unified — Gemma4UnifiedProcessor (modelo Gemma4UnifiedConfig)
• git — GitProcessor (modelo GitConfig)
• glm46v — Glm46VProcessor (modelo Glm46VConfig)
• glm4v — Glm4vProcessor (modelo Glm4vConfig)
• glm4v_moe — Glm4vProcessor (modelo Glm4vMoeConfig)
• glm5_next — Glm5NextProcessor (modelo Glm5NextConfig)
• glm_image — GlmImageProcessor (modelo GlmImageConfig)
• glmasr — GlmAsrProcessor (modelo GlmAsrConfig)
• got_ocr2 — GotOcr2Processor (modelo GotOcr2Config)
• gradium_phonon_tts — PocketTTSProcessor (modelo GradiumPhononTTSConfig)
• granite4_vision — Granite4VisionProcessor (modelo Granite4VisionConfig)
• granite_speech — GraniteSpeechProcessor (modelo GraniteSpeechConfig)
• granite_speech5_ctc — GraniteSpeech5Processor (modelo GraniteSpeech5CTCConfig)
• granite_speech_plus — GraniteSpeechProcessor (modelo GraniteSpeechPlusConfig)
• grounding-dino — GroundingDinoProcessor (modelo GroundingDinoConfig)
• groupvit — CLIPProcessor (modelo GroupViTConfig)
• higgs_audio_v2 — HiggsAudioV2Processor (modelo HiggsAudioV2Config)
• hubert — Wav2Vec2Processor (modelo HubertConfig)
• hunyuan_vl — HunYuanVLProcessor (modelo HunYuanVLConfig)
• hyperclovax_vision_v2 — HyperCLOVAXVisionV2Processor (modelo HyperCLOVAXVisionV2Config)
• idefics — IdeficsProcessor (modelo IdeficsConfig)
• idefics2 — Idefics2Processor (modelo Idefics2Config)
• idefics3 — Idefics3Processor (modelo Idefics3Config)
• inkling_mm_model — InklingProcessor (modelo InklingConfig)
• instructblip — InstructBlipProcessor (modelo InstructBlipConfig)
• instructblipvideo — InstructBlipVideoProcessor (modelo InstructBlipVideoConfig)
• internvl — InternVLProcessor (modelo InternVLConfig)
• janus — JanusProcessor (modelo JanusConfig)
• kimi_k25 — Kimi_K25Processor (modelo Kimi_K25Config)
• kyutai_speech_to_text — KyutaiSpeechToTextProcessor (modelo KyutaiSpeechToTextConfig)
• lasr_ctc — LasrProcessor (modelo LasrCTCConfig)
• lasr_encoder — LasrProcessor (modelo LasrEncoderConfig)
• layoutlmv2 — LayoutLMv2Processor (modelo LayoutLMv2Config)
• layoutlmv3 — LayoutLMv3Processor (modelo LayoutLMv3Config)
• layoutxlm — LayoutXLMProcessor (modelo LayoutXLMConfig)
• lfm2_vl — Lfm2VlProcessor (modelo Lfm2VlConfig)
• lighton_ocr — LightOnOcrProcessor (modelo LightOnOcrConfig)
• llama4 — Llama4Processor (modelo Llama4Config)
• llava — LlavaProcessor (modelo LlavaConfig)
• llava_next — LlavaNextProcessor (modelo LlavaNextConfig)
• llava_next_video — LlavaNextVideoProcessor (modelo LlavaNextVideoConfig)
• llava_onevision — LlavaOnevisionProcessor (modelo LlavaOnevisionConfig)
• markuplm — MarkupLMProcessor (modelo MarkupLMConfig)
• metaclip_2 — CLIPProcessor (modelo MetaClip2Config)
• mgp-str — MgpstrProcessor (modelo MgpstrConfig)
• minicpm_ocr — MiniCPMOcrProcessor (modelo MiniCPMOcrConfig)
• minicpmv4_6 — MiniCPMV4_6Processor (modelo MiniCPMV4_6Config)
• minicpmv4_7 — MiniCPMV4_7Processor (modelo MiniCPMV4_7Config)
• minimax_m3_vl — MiniMaxM3VLProcessor (modelo MiniMaxM3VLConfig)
• mistral3 — PixtralProcessor (modelo Mistral3Config)
• mistral_large4 — PixtralProcessor (modelo MistralLarge4Config)
• mllama — MllamaProcessor (modelo MllamaConfig)
• mm-grounding-dino — GroundingDinoProcessor (modelo MMGroundingDinoConfig)
• modernvbert — Idefics3Processor (modelo ModernVBertConfig)
• molmo2 — Molmo2Processor (modelo Molmo2Config)
• moonshine — Wav2Vec2Processor (modelo MoonshineConfig)
• moonshine_streaming — MoonshineStreamingProcessor (modelo MoonshineStreamingConfig)
• moss_transcribe_diarize — MossTranscribeDiarizeProcessor (modelo MossTranscribeDiarizeConfig)
• muse_glimmer — MuseGlimmerProcessor (modelo MuseGlimmerConfig)
• muse_spark — MuseSparkProcessor (modelo MuseSparkConfig)
• musicflamingo — MusicFlamingoProcessor (modelo MusicFlamingoConfig)
• musicgen — MusicgenProcessor (modelo MusicgenConfig)
• musicgen_melody — MusicgenMelodyProcessor (modelo MusicgenMelodyConfig)
• nemo_speechlm — Nemotron3_5TranscribeProcessor (modelo Nemotron3_5TranscribeConfig)
• nemotron3_5_asr — Nemotron3_5AsrProcessor (modelo Nemotron3_5AsrConfig)
• nemotron3_diarization — Nemotron3DiarizationProcessor (modelo Nemotron3DiarizationConfig)
• nemotron_asr_streaming — NemotronAsrStreamingProcessor (modelo NemotronAsrStreamingConfig)
• nemotron_h_omni — NemotronH_Omni_Reasoning_V3Processor (modelo NemotronH_Omni_Reasoning_V3_Config)
• neomme — NeoMMEProcessor (modelo NeoMMEConfig)
• nougat — NougatProcessor (modelo NougatConfig)
• omdet-turbo — OmDetTurboProcessor (modelo OmDetTurboConfig)
• oneformer — OneFormerProcessor (modelo OneFormerConfig)
• ovis2 — Ovis2Processor (modelo Ovis2Config)
• owlv2 — Owlv2Processor (modelo Owlv2Config)
• owlvit — OwlViTProcessor (modelo OwlViTConfig)
• paddleocr_vl — PaddleOCRVLProcessor (modelo PaddleOCRVLConfig)
• paligemma — PaliGemmaProcessor (modelo PaliGemmaConfig)
• parakeet_ctc — ParakeetProcessor (modelo ParakeetCTCConfig)
• parakeet_rnnt — ParakeetProcessor (modelo ParakeetRNNTConfig)
• parakeet_tdt — ParakeetProcessor (modelo ParakeetTDTConfig)
• pe_audio — PeAudioProcessor (modelo PeAudioConfig)
• pe_audio_video — PeAudioVideoProcessor (modelo PeAudioVideoConfig)
• pe_video — PeVideoProcessor (modelo PeVideoConfig)
• perception_lm — PerceptionLMProcessor (modelo PerceptionLMConfig)
• phi4_multimodal — Phi4MultimodalProcessor (modelo Phi4MultimodalConfig)
• pi0 — PI0Processor (modelo PI0Config)
• pix2struct — Pix2StructProcessor (modelo Pix2StructConfig)
• pixtral — PixtralProcessor (modelo PixtralVisionConfig)
• pocket_tts — PocketTTSProcessor (modelo PocketTTSConfig)
• pop2piano — Pop2PianoProcessor (modelo Pop2PianoConfig)
• pp_chart2table — PPChart2TableProcessor (modelo PPChart2TableConfig)
• pp_formulanet — PPFormulaNetProcessor (modelo PPFormulaNetConfig)
• qianfan_ocr — QianfanOCRProcessor (modelo QianfanOCRConfig)
• qwen2_5_omni — Qwen2_5OmniProcessor (modelo Qwen2_5OmniConfig)
• qwen2_5_vl — Qwen2_5_VLProcessor (modelo Qwen2_5_VLConfig)
• qwen2_audio — Qwen2AudioProcessor (modelo Qwen2AudioConfig)
• qwen2_vl — Qwen2VLProcessor (modelo Qwen2VLConfig)
• qwen3_5 — Qwen3VLProcessor (modelo Qwen3_5Config)
• qwen3_5_moe — Qwen3VLProcessor (modelo Qwen3_5MoeConfig)
• qwen3_asr — Qwen3ASRProcessor (modelo Qwen3ASRConfig)
• qwen3_omni_moe — Qwen3OmniMoeProcessor (modelo Qwen3OmniMoeConfig)
• qwen3_vl — Qwen3VLProcessor (modelo Qwen3VLConfig)
• qwen3_vl_moe — Qwen3VLProcessor (modelo Qwen3VLMoeConfig)
• qwen4_exp — Qwen3VLProcessor (modelo Qwen4ExpConfig)
• sam — SamProcessor (modelo SamConfig)
• sam2 — Sam2Processor (modelo Sam2Config)
• sam2_video — Sam2VideoProcessor (modelo Sam2VideoConfig)
• sam3 — Sam3Processor (modelo Sam3Config)
• sam3_lite_text — Sam3Processor (modelo Sam3LiteTextConfig)
• sam3_tracker — Sam3TrackerProcessor (modelo Sam3TrackerConfig)
• sam3_tracker_video — Sam3TrackerVideoProcessor (modelo Sam3TrackerVideoConfig)
• sam3_video — Sam3VideoProcessor (modelo Sam3VideoConfig)
• sam_hq — SamHQProcessor (modelo SamHQConfig)
• seamless_m4t — SeamlessM4TProcessor (modelo SeamlessM4TConfig)
• sew — Wav2Vec2Processor (modelo SEWConfig)
• sew-d — Wav2Vec2Processor (modelo SEWDConfig)
• siglip — SiglipProcessor (modelo SiglipConfig)
• siglip2 — Siglip2Processor (modelo Siglip2Config)
• smolvlm — SmolVLMProcessor (modelo SmolVLMConfig)
• speech_to_text — Speech2TextProcessor (modelo Speech2TextConfig)
• speecht5 — SpeechT5Processor (modelo SpeechT5Config)
• step3p7 — Step3p7Processor (modelo Step3p7Config)
• t5gemma2 — Gemma3Processor (modelo T5Gemma2Config)
• t5gemma2_encoder — Gemma3Processor (modelo T5Gemma2EncoderConfig)
• tipsv2 — Tipsv2Processor (modelo Tipsv2Config)
• trocr — TrOCRProcessor (modelo TrOCRConfig)
• tvp — TvpProcessor (modelo TvpConfig)
• udop — UdopProcessor (modelo UdopConfig)
• ultravbert — Idefics3Processor (modelo UltraVBertConfig)
• unispeech — Wav2Vec2Processor (modelo UniSpeechConfig)
• unispeech-sat — Wav2Vec2Processor (modelo UniSpeechSatConfig)
• unlimited_ocr — UnlimitedOcrProcessor (modelo UnlimitedOcrConfig)
• vibevoice — VibeVoiceProcessor (modelo VibeVoiceConfig)
• vibevoice_asr — VibeVoiceAsrProcessor (modelo VibeVoiceAsrConfig)
• video_llama_3 — VideoLlama3Processor (modelo VideoLlama3Config)
• video_llava — VideoLlavaProcessor (modelo VideoLlavaConfig)
• videoprism — VideoPrismProcessor (modelo VideoPrismConfig)
• vilt — ViltProcessor (modelo ViltConfig)
• vipllava — LlavaProcessor (modelo VipLlavaConfig)
• vision-text-dual-encoder — VisionTextDualEncoderProcessor (modelo VisionTextDualEncoderConfig)
• voxtral — VoxtralProcessor (modelo VoxtralConfig)
• voxtral_realtime — VoxtralRealtimeProcessor (modelo VoxtralRealtimeConfig)
• wav2vec2 — Wav2Vec2Processor (modelo Wav2Vec2Config)
• wav2vec2-bert — Wav2Vec2BertProcessor (modelo Wav2Vec2BertConfig)
• wav2vec2-conformer — Wav2Vec2Processor (modelo Wav2Vec2ConformerConfig)
• wavlm — Wav2Vec2Processor (modelo WavLMConfig)
• whisper — WhisperProcessor (modelo WhisperConfig)

É necessário passar token=True quando você deseja usar um modelo privado.

>>> from transformers import AutoProcessor

>>> # Download processor from huggingface.co and cache.
>>> processor = AutoProcessor.from_pretrained("facebook/wav2vec2-base-960h")

>>> # If processor files are in a directory (e.g. processor was saved using *save_pretrained('./test/saved_model/')*)
>>> # processor = AutoProcessor.from_pretrained("./test/saved_model/")

Registre um novo processador para esta classe.