T^(h/dZddlmZddlmZmZmZmZddlm Z ddl m Z ddl m Z ddlmZdd lmZdd lmZmZej*eZGd d e ZGd de Zd dgZy)zPerceiver model configuration) OrderedDict)AnyMappingOptionalUnion)PretrainedConfig)FeatureExtractionMixin) OnnxConfig) compute_effective_axis_dimension)PreTrainedTokenizerBase) TensorTypeloggingcfeZdZdZdZdddddddd d d ddd d d dddddddgdddgdddffd ZxZS)PerceiverConfiga This is the configuration class to store the configuration of a [`PerceiverModel`]. It is used to instantiate an Perceiver model according to the specified arguments, defining the model architecture. Instantiating a configuration with the defaults will yield a similar configuration to that of the Perceiver [deepmind/language-perceiver](https://huggingface.co/deepmind/language-perceiver) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: num_latents (`int`, *optional*, defaults to 256): The number of latents. d_latents (`int`, *optional*, defaults to 1280): Dimension of the latent embeddings. d_model (`int`, *optional*, defaults to 768): Dimension of the inputs. Should only be provided in case [*PerceiverTextPreprocessor*] is used or no preprocessor is provided. num_blocks (`int`, *optional*, defaults to 1): Number of blocks in the Transformer encoder. num_self_attends_per_block (`int`, *optional*, defaults to 26): The number of self-attention layers per block. num_self_attention_heads (`int`, *optional*, defaults to 8): Number of attention heads for each self-attention layer in the Transformer encoder. num_cross_attention_heads (`int`, *optional*, defaults to 8): Number of attention heads for each cross-attention layer in the Transformer encoder. qk_channels (`int`, *optional*): Dimension to project the queries + keys before applying attention in the cross-attention and self-attention layers of the encoder. Will default to preserving the dimension of the queries if not specified. v_channels (`int`, *optional*): Dimension to project the values before applying attention in the cross-attention and self-attention layers of the encoder. Will default to preserving the dimension of the queries if not specified. cross_attention_shape_for_attention (`str`, *optional*, defaults to `"kv"`): Dimension to use when downsampling the queries and keys in the cross-attention layer of the encoder. self_attention_widening_factor (`int`, *optional*, defaults to 1): Dimension of the feed-forward layer in the cross-attention layer of the Transformer encoder. cross_attention_widening_factor (`int`, *optional*, defaults to 1): Dimension of the feed-forward layer in the self-attention layers of the Transformer encoder. hidden_act (`str` or `function`, *optional*, defaults to `"gelu"`): The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`, `"relu"`, `"selu"` and `"gelu_new"` are supported. attention_probs_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout ratio for the attention probabilities. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. layer_norm_eps (`float`, *optional*, defaults to 1e-12): The epsilon used by the layer normalization layers. use_query_residual (`float`, *optional*, defaults to `True`): Whether to add a query residual in the cross-attention layer of the encoder. vocab_size (`int`, *optional*, defaults to 262): Vocabulary size for the masked language modeling model. max_position_embeddings (`int`, *optional*, defaults to 2048): The maximum sequence length that the masked language modeling model might ever be used with. Typically set this to something large just in case (e.g., 512 or 1024 or 2048). image_size (`int`, *optional*, defaults to 56): Size of the images after preprocessing, for [`PerceiverForImageClassificationLearned`]. train_size (`List[int]`, *optional*, defaults to `[368, 496]`): Training size of the images for the optical flow model. num_frames (`int`, *optional*, defaults to 16): Number of video frames used for the multimodal autoencoding model. audio_samples_per_frame (`int`, *optional*, defaults to 1920): Number of audio samples per frame for the multimodal autoencoding model. samples_per_patch (`int`, *optional*, defaults to 16): Number of audio samples per patch when preprocessing the audio for the multimodal autoencoding model. output_shape (`List[int]`, *optional*, defaults to `[1, 16, 224, 224]`): Shape of the output (batch_size, num_frames, height, width) for the video decoder queries of the multimodal autoencoding model. This excludes the channel dimension. output_num_channels (`int`, *optional*, defaults to 512): Number of output channels for each modalitiy decoder. Example: ```python >>> from transformers import PerceiverModel, PerceiverConfig >>> # Initializing a Perceiver deepmind/language-perceiver style configuration >>> configuration = PerceiverConfig() >>> # Initializing a model from the deepmind/language-perceiver style configuration >>> model = PerceiverModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` perceiveriiNkvgelug?g{Gz?g-q=Tii8ipii)rrriic t|di|||_||_||_||_||_||_||_||_ | |_ | |_ | |_ | |_ | |_||_||_||_||_||_||_||_||_||_||_||_||_||_||_y)N)super__init__ num_latents d_latentsd_model num_blocksnum_self_attends_per_blocknum_self_attention_headsnum_cross_attention_heads qk_channels v_channels#cross_attention_shape_for_attentionself_attention_widening_factorcross_attention_widening_factor hidden_actattention_probs_dropout_probinitializer_rangelayer_norm_epsuse_query_residual vocab_sizemax_position_embeddings image_size train_size num_framesaudio_samples_per_framesamples_per_patch output_shapeoutput_num_channels_label_trainable_num_channels)selfr r!r"r#r$r%r&r'r(r)r*r+r,r-r.r/r0r1r2r3r4r5r6r7r8r9r:kwargs __class__s s/var/www/skyplay_api_hub/venv/lib/python3.12/site-packages/transformers/models/perceiver/configuration_perceiver.pyrzPerceiverConfig.__init__vs> "6"&" $*D'(@%)B&&$3V0.L+/N,$,H)!2,"4$'>$$$$'>$!2(#6 -J*)__name__ __module__ __qualname____doc__ model_typer __classcell__)r=s@r>rrstR hJ#%!""#,0'(()%( $: $&&*9?K?Kr?rceZdZedeeeeefffdZedefdZ dde ddeded ed e d e e d ed ededeeeffdZy)PerceiverOnnxConfigreturncZ|jdk(rdddd}nddd}td|fd|fgS) Nzmultiple-choicebatchchoicesequence)rr)rrinputsattention_mask)taskr)r; dynamic_axiss r>rNzPerceiverOnnxConfig.inputssG 99) )&8 CL&:6L<(!<0   r?cy)Ng-C6?r)r;s r>atol_for_validationz'PerceiverOnnxConfig.atol_for_validationsr?N preprocessor)r r batch_size seq_length num_choicesis_pair framework num_channels image_width image_heightc Bt|trt|tjd}|j |} t|tj | }djdg|zg|z} t|| |} | jd| d<| St|trj|jddk(rXt|tj }|j||| |} t|| | } | jd| d<| Std ) Nr)fixed_dimensionnum_token_to_add a)return_tensors input_idsrN pixel_values)r^)imagesrbz\Unable to generate dummy inputs for the model. Please provide a tokenizer or a preprocessor.) isinstancer r r default_fixed_batchnum_special_tokens_to_adddefault_fixed_sequencejoindictpopr model_input_names_generate_dummy_images ValueError) r;rTrUrVrWrXrYrZr[r\ token_to_add dummy_inputrNs r>generate_dummy_inputsz)PerceiverOnnxConfig.generate_dummy_inputss l$; <9J,J,J]^J(AA'JL9J,M,M`lJ88SE?Z78:EK,{9MNF%zz+6F8 M  &< =,B`B`abBcguBu9*V`VtVtuJ55j,P\^ijK,k)TUF%zz.9F8 Mn r?)rsrsFNr(rt)r@rArBpropertyrstrintrNfloatrSrboolrrrrrrr?r>rGrGs   WS#X%6 67    U *.'OP'' '  '  'J''''' c 'r?rGN)rC collectionsrtypingrrrrconfiguration_utilsr feature_extraction_utilsr onnxr onnx.utilsr tokenization_utils_baser utilsrr get_loggerr@loggerrrG__all__rr?r>rse$#003>:>(   H %VK&VKr9*9x 3 4r?