^j !dZddlmZddlmZddlmZmZeje Z edeGdd eZ edeGd d eZ edeGd d eZ gdZy)zALIGN model configuration)strict)PreTrainedConfig)auto_docstringloggingzkakaobrain/align-base) checkpointceZdZUdZdZdZdZeed<dZ eed<dZ eed <dZ eed <d Z eed <d Z eed<dZeezed<dZeezed<dZeed<dZeed<dZeed<dZeed<dZedzed<dZedzed<dZeeezdzed<y)AlignTextConfiga Example: ```python >>> from transformers import AlignTextConfig, AlignTextModel >>> # Initializing a AlignTextConfig with kakaobrain/align-base style configuration >>> configuration = AlignTextConfig() >>> # Initializing a AlignTextModel (with random weights) from the kakaobrain/align-base style configuration >>> model = AlignTextModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```align_text_model text_configi:w vocab_sizei hidden_size num_hidden_layersnum_attention_headsi intermediate_sizegelu hidden_actg?hidden_dropout_probattention_probs_dropout_probimax_position_embeddingstype_vocab_size{Gz?initializer_rangeg-q=layer_norm_epsrN pad_token_id bos_token_id eos_token_id)__name__ __module__ __qualname____doc__ model_typebase_config_keyr int__annotations__rrrrrstrrfloatrrrrrrrrlistx/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/transformers/models/align/configuration_align.pyr r s $J#OJKs!!!s!J'**03 %#+3#&S&OS#u#!NE! L#* #L#*#+/L#S /D(/r,r ceZdZUdZdZdZdZeed<dZ ee eze eefzed<dZ e ed <d Ze ed <d Zeed <dZe ee edfzed<dZe ee edfzed<dZe ee edfzed<dZe e edfzed<dZe ee edfzed<dZe ee edfzed<dZe ee edfzed<dZe ed<dZeed <d!Zeed"<d#Zeed$<d%Ze ed&<d'Ze ed(<d)Ze ed*<d+Ze ezed,<fd-Z xZ!S).AlignVisionConfiga width_coefficient (`float`, *optional*, defaults to 2.0): Scaling coefficient for network width at each stage. depth_coefficient (`float`, *optional*, defaults to 3.1): Scaling coefficient for network depth at each stage. depth_divisor (`int`, *optional*, defaults to 8): A unit of network width. kernel_sizes (`list[int]`, *optional*, defaults to `[3, 3, 5, 3, 5, 5, 3]`): List of kernel sizes to be used in each block. in_channels (`list[int]`, *optional*, defaults to `[32, 16, 24, 40, 80, 112, 192]`): List of input channel sizes to be used in each block for convolutional layers. out_channels (`list[int]`, *optional*, defaults to `[16, 24, 40, 80, 112, 192, 320]`): List of output channel sizes to be used in each block for convolutional layers. depthwise_padding (`list[int]`, *optional*, defaults to `[]`): List of block indices with square padding. strides (`list[int]`, *optional*, defaults to `[1, 2, 2, 2, 1, 2, 1]`): List of stride sizes to be used in each block for convolutional layers. num_block_repeats (`list[int]`, *optional*, defaults to `[1, 2, 2, 3, 3, 4, 1]`): List of the number of times each block is to repeated. expand_ratios (`list[int]`, *optional*, defaults to `[1, 6, 6, 6, 6, 6, 6]`): List of scaling coefficient of each block. squeeze_expansion_ratio (`float`, *optional*, defaults to 0.25): Squeeze expansion ratio. hidden_dim (`int`, *optional*, defaults to 1280): The hidden dimension of the layer before the classification head. pooling_type (`str` or `function`, *optional*, defaults to `"mean"`): Type of final pooling to be applied before the dense classification head. Available options are [`"mean"`, `"max"`] batch_norm_momentum (`float`, *optional*, defaults to 0.99): The momentum used by the batch normalization layers. drop_connect_rate (`float`, *optional*, defaults to 0.2): The drop rate for skip connections. Example: ```python >>> from transformers import AlignVisionConfig, AlignVisionModel >>> # Initializing a AlignVisionConfig with kakaobrain/align-base style configuration >>> configuration = AlignVisionConfig() >>> # Initializing a AlignVisionModel (with random weights) from the kakaobrain/align-base style configuration >>> model = AlignVisionModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```align_vision_model vision_configr num_channelsiX image_sizeg@width_coefficientg@depth_coefficient depth_divisor)rrrr8r8r. kernel_sizes) (Pp in_channels)r;r<r=r>r?r@i@ out_channelsr+depthwise_padding)rrrrDrrDstrides)rDrrrrrDnum_block_repeats)rDrHrHrHrHrH expand_ratiosg?squeeze_expansion_ratioswishri hidden_dimmean pooling_typerrgMbP?batch_norm_epsgGz?batch_norm_momentumg?drop_connect_ratec t|jdz|_dD]"}t||t t ||$t |di|y)NrF)r9rArBrCrErGrIr+)sumrGrsetattrr*getattrsuper __post_init__)selfkwargsattr __class__s r-rWzAlignVisionConfig.__post_init__sX!$T%;%;!>> from transformers import AlignConfig, AlignModel >>> # Initializing a AlignConfig with kakaobrain/align-base style configuration >>> configuration = AlignConfig() >>> # Initializing a AlignModel (with random weights) from the kakaobrain/align-base style configuration >>> model = AlignModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config >>> # We can also initialize a AlignConfig from a AlignTextConfig and a AlignVisionConfig >>> from transformers import AlignTextConfig, AlignVisionConfig >>> # Initializing ALIGN Text and Vision configurations >>> config_text = AlignTextConfig() >>> config_vision = AlignVisionConfig() >>> config = AlignConfig(text_config=config_text, vision_config=config_vision) ```align)r r1Nr r1iprojection_dimg?temperature_init_valuerrc |j%t|_tjdn4t |jt rtdi|j|_|j %t|_tjdn4t |j t rtdi|j |_t|$di|y)NzP`text_config` is `None`. Initializing the `AlignTextConfig` with default values.zT`vision_config` is `None`. initializing the `AlignVisionConfig` with default values.r+) r r loggerinfo isinstancedictr1r/rVrW)rXrYr[s r-rWzAlignConfig.__post_init__s    #.0D  KKj k (($ /.B1A1ABD     %!2!4D  KKn o **D 1!2!HT5G5G!HD  ''r,)r r!r"r#r$r r/ sub_configsr rhrr'r1rbr&rcr)rrWr]r^s@r-r`r`sn 8J"1DUVK26K((4/648M4**T18NC$'E'#u# ( (r,r`)r r/r`N)r#huggingface_hub.dataclassesrconfiguration_utilsrutilsrr get_loggerr rer r/r`__all__r+r,r-ros .3,   H %23"0&"04"0J23V((V(4V(r233("3(43(l Br,