U 4Af @sdZddlZddlmZmZmZddlmZddlm Z e e Z GdddeZ Gd d d eZGd d d eZGd ddeZGdddeZdS)zFLAVA model configurationsN)AnyDictUnion)PretrainedConfig)loggingcsfeZdZdZdZdeeeeeeeeeeeeeeedfdd Ze e e e j fddddZZS)FlavaImageConfiga This is the configuration class to store the configuration of a [`FlavaImageModel`]. It is used to instantiate an FLAVA 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 FLAVA [facebook/flava-full](https://huggingface.co/facebook/flava-full) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: hidden_size (`int`, *optional*, defaults to 768): Dimensionality of the encoder layers and the pooler layer. num_hidden_layers (`int`, *optional*, defaults to 12): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 12): Number of attention heads for each attention layer in the Transformer encoder. intermediate_size (`int`, *optional*, defaults to 3072): Dimensionality of the "intermediate" (i.e., feed-forward) layer in 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. hidden_dropout_prob (`float`, *optional*, defaults to 0.0): The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. attention_probs_dropout_prob (`float`, *optional*, defaults to 0.0): 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. image_size (`int`, *optional*, defaults to 224): The size (resolution) of each image. patch_size (`int`, *optional*, defaults to 16): The size (resolution) of each patch. num_channels (`int`, *optional*, defaults to 3): The number of input channels. qkv_bias (`bool`, *optional*, defaults to `True`): Whether to add a bias to the queries, keys and values. mask_token (`bool`, *optional*, defaults to `True`): Whether to use a mask token or not. Used in MIM (Masked Image Modeling) loss for FLAVA. vocab_size (`int`, *optional*, defaults to 8192): Vocabulary size of the [`FlavaImageCodebook`] used in conjunction with [`FlavaImageModel`] for MIM (Masked Image Modeling) loss for FLAVA. Example: ```python >>> from transformers import FlavaImageConfig, FlavaImageModel >>> # Initializing a FlavaImageModel with style configuration >>> configuration = FlavaImageConfig() >>> # Initializing a FlavaImageModel model (with random weights) from the style configuration >>> model = FlavaImageModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```Zflava_image_model gelu{Gz?-q=rT ) hidden_sizenum_hidden_layersnum_attention_headsintermediate_size hidden_acthidden_dropout_probattention_probs_dropout_probinitializer_rangelayer_norm_eps image_size patch_size num_channelsqkv_bias mask_token vocab_sizec sltjf|||_||_||_||_||_||_||_||_ | |_ | |_ | |_ | |_ | |_||_||_dSN)super__init__rrrrrrrrrrrrrr r!)selfrrrrrrrrrrrrrr r!kwargs __class__Q/tmp/pip-unpacked-wheel-zw5xktn0/transformers/models/flava/configuration_flava.pyr$Zs zFlavaImageConfig.__init__rpretrained_model_name_or_pathreturncKs~|||j|f|\}}|ddkr2|d}d|krpt|drp|d|jkrptd|dd|jd|j|f|S)N model_typeflava image_configYou are using a model of type to instantiate a model of type N. This is not supported for all configurations of models and can yield errors.Z_set_token_in_kwargsZget_config_dictgethasattrr.loggerwarning from_dictclsr,r&Z config_dictr)r)r*from_pretraineds  z FlavaImageConfig.from_pretrained)r r r r r r r rrrrrTTr__name__ __module__ __qualname____doc__r.intfloatboolr$ classmethodrstrosPathLiker< __classcell__r)r)r'r*rsH<%rcsfeZdZdZdZdeeeeeeeeeeeeeeedfdd Z e e ee j fddddZZS)FlavaTextConfigaC This is the configuration class to store the configuration of a [`FlavaTextModel`]. It is used to instantiate an FLAVA 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 FLAVA [facebook/flava-full](https://huggingface.co/facebook/flava-full) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: vocab_size (`int`, *optional*, defaults to 30522): Vocabulary size of the BERT model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`FlavaTextModel`]. type_vocab_size (`int`, *optional*, defaults to 2): The vocabulary size of the `token_type_ids` passed when calling [`FlavaTextModel`]. Note that even though text encoder allows `token_type_ids`'s value as 2, for text-only pretraining and fine-tuning, only 1 is used similar to RoBERTa. max_position_embeddings (`int`, *optional*, defaults to 512): The maximum sequence length that this model might ever be used with. Typically set this to something large just in case (e.g., 512 or 1024 or 2048). For VL, max_length passed to model is 77. position_embedding_type (`str`, *optional*, defaults to `"absolute"`): Type of position embedding. Choose one of `"absolute"`, `"relative_key"`, `"relative_key_query"`. For positional embeddings use `"absolute"`. For more information on `"relative_key"`, please refer to [Self-Attention with Relative Position Representations (Shaw et al.)](https://arxiv.org/abs/1803.02155). For more information on `"relative_key_query"`, please refer to *Method 4* in [Improve Transformer Models with Better Relative Position Embeddings (Huang et al.)](https://arxiv.org/abs/2009.13658). hidden_size (`int`, *optional*, defaults to 768): Dimensionality of the encoder layers and the pooler layer. num_hidden_layers (`int`, *optional*, defaults to 12): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 12): Number of attention heads for each attention layer in the Transformer encoder. intermediate_size (`int`, *optional*, defaults to 3072): Dimensionality of the "intermediate" (i.e., feed-forward) layer in 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. hidden_dropout_prob (`float`, *optional*, defaults to 0.1): The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. 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. image_size (`int`, *optional*, defaults to 224): The size (resolution) of each image. patch_size (`int`, *optional*, defaults to 16): The size (resolution) of each patch. num_channels (`int`, *optional*, defaults to 3): The number of input channels. qkv_bias (`bool`, *optional*, defaults to `True`): Whether to add a bias to the queries, keys and values. Example: ```python >>> from transformers import FlavaTextConfig, FlavaTextModel >>> # Initializing a FlavaTextModel with style configuration >>> configuration = FlavaTextConfig() >>> # Initializing a FlavaTextModel model (with random weights) from the style configuration >>> model = FlavaTextModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```Zflava_text_model:wabsoluter r r r r rrrT)r!type_vocab_sizemax_position_embeddingsposition_embedding_typerrrrrrrrr pad_token_idrc sltjf|||_||_||_||_||_||_||_||_ | |_ | |_ | |_ | |_ | |_||_||_dSr")r#r$r!rOrPrQrrrrrrrrrrrR)r%r!rOrPrQrrrrrrrrrrRrr&r'r)r*r$s zFlavaTextConfig.__init__rr+cKs~|||j|f|\}}|ddkr2|d}d|krpt|drp|d|jkrptd|dd|jd|j|f|S)Nr.r/ text_configr1r2r3r4r:r)r)r*r<s  zFlavaTextConfig.from_pretrained)rKrLrMrNr r r r r r r rrrT)r>r?r@rAr.rBrFrCrDr$rErrGrHr<rIr)r)r'r*rJsHG%rJc s^eZdZdZdZdeeeeeeeeeeed fd d Ze e e e j fddddZZS)FlavaMultimodalConfiga This is the configuration class to store the configuration of a [`FlavaMultimodalModel`]. It is used to instantiate an FLAVA 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 FLAVA [facebook/flava-full](https://huggingface.co/facebook/flava-full) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: hidden_size (`int`, *optional*, defaults to 768): Dimensionality of the encoder layers and the pooler layer. num_hidden_layers (`int`, *optional*, defaults to 6): Number of hidden layers in the Transformer encoder. num_attention_heads (`int`, *optional*, defaults to 12): Number of attention heads for each attention layer in the Transformer encoder. intermediate_size (`int`, *optional*, defaults to 3072): Dimensionality of the "intermediate" (i.e., feed-forward) layer in 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. hidden_dropout_prob (`float`, *optional*, defaults to 0.0): The dropout probability for all fully connected layers in the embeddings, encoder, and pooler. attention_probs_dropout_prob (`float`, *optional*, defaults to 0.0): 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. qkv_bias (`bool`, *optional*, defaults to `True`): Whether to add a bias to the queries, keys and values. use_cls_token (`bool`, *optional*, defaults to `True`): Whether to use an extra CLS token for multimodal settings. Usually needed by the FLAVA model. Example: ```python >>> from transformers import FlavaMultimodalConfig, FlavaMultimodalModel >>> # Initializing a FlavaMultimodalModel with style configuration >>> configuration = FlavaMultimodalConfig() >>> # Initializing a FlavaMultimodalModel model (with random weights) from the style configuration >>> model = FlavaMultimodalModel(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```Zflava_multimodal_modelr r r r r rrT) rrrrrrrrrr use_cls_tokenc  sTtjf| ||_||_||_||_||_||_||_||_ | |_ | |_ | |_ dSr") r#r$rrrrrrrrrrrV) r%rrrrrrrrrrrVr&r'r)r*r$KszFlavaMultimodalConfig.__init__rr+cKs~|||j|f|\}}|ddkr2|d}d|krpt|drp|d|jkrptd|dd|jd|j|f|S)Nr.r/multimodal_configr1r2r3r4r:r)r)r*r<hs  z%FlavaMultimodalConfig.from_pretrained) r rUr r r r r rrTTr=r)r)r'r*rTs84rTc sReZdZdZdeeeeeeed fd d Zeee e j fd d ddZ Z S)FlavaImageCodebookConfigZflava_image_codebookrrLrTr) num_groupsinput_channelsnum_blocks_per_grouprr!freezerc  s<tjf|||_||_||_||_||_||_||_dSr") r#r$r[r\r]rr!r^r) r%r[r\r]rr!r^rr&r'r)r*r$s z!FlavaImageCodebookConfig.__init__rr+cKs~|||j|f|\}}|ddkr2|d}d|krpt|drp|d|jkrptd|dd|jd|j|f|S)Nr.r/image_codebook_configr1r2r3r4r:r)r)r*r<s  z(FlavaImageCodebookConfig.from_pretrained)rYrrLrZrTr)r>r?r@r.rBrCr$rErrFrGrHr<rIr)r)r'r*rX{s&/rXcseZdZdZdZdeeefeeefeeefeeefee ee e e ee e e e e e e e e d fd d Z e e eeedddZZS) FlavaConfiga [`FlavaConfig`] is the configuration class to store the configuration of a [`FlavaModel`]. It is used to instantiate FLAVA model according to the specified arguments, defining the text model, image model, image codebook and multimodal model configs. Instantiating a configuration with the defaults will yield a similar configuration to that of the FLAVA [facebook/flava-full](https://huggingface.co/facebook/flava-full) architecture. Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the documentation from [`PretrainedConfig`] for more information. Args: text_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`FlavaTextConfig`]. image_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`FlavaImageConfig`]. multimodal_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`FlavaMultimodalConfig`]. hidden_size (`int`, *optional*, defaults to 768): Dimensionality of the encoder layers and the pooler layer. layer_norm_eps (`float`, *optional*, defaults to 1e-12): The epsilon used by the layer normalization layers. projection_dim (`int`, *optional*, defaults to 512): Dimensionality of text and image projection layers. logit_scale_init_value (`float`, *optional*, defaults to 2.6592): The initial value of the *logit_scale* parameter. Default is used as per the original FLAVA/CLIP implementation. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. ce_ignore_index (`int`, *optional*, defaults to -100): Cross entropy index to ignore. mim_weight (`float`, *optional*, defaults to 1.0): Weight to be assigned to MIM (Masked Image Modeling) unimodal loss mlm_weight (`float`, *optional*, defaults to 1.0): Weight to be assigned to MLM (Masked Language Modeling) unimodal loss global_contrastive_weight (`float`, *optional*, defaults to 1.0): Weight to be assigned to global contrastive cross-alignment loss. itm_weight (`float`, *optional*, defaults to 1.0): Weight to be assigned to image-text matching multimodal loss. mmm_image_weight (`float`, *optional*, defaults to 1.0): Weight to be assigned to MMM loss's image part. mmm_text_weight (`float`, *optional*, defaults to 1.0): Weight to be assigned to MMM loss's text part. global_backprop_contrastive (`bool`, *optional*, defaults to `True`): Whether to use global backpropgation through all workers in contrastive loss. skip_unmasked_multimodal_encoder (`bool`, *optional*, defaults to `True`): Whether to skip running unmasked multimodal encoder whose outputs are not used by FLAVA losses. return_loss (`bool`, *optional*, defaults to `True`): Whether to return loss or not kwargs (*optional*): Dictionary of keyword arguments. Example: ```python >>> from transformers import FlavaConfig, FlavaModel, FlavaForPreTraining >>> # Initializing a FlavaConfig with style configuration >>> configuration = FlavaConfig() >>> # Initializing a FlavaModel and FlavaForPreTraining model (with random weights) from the style configuration >>> model = FlavaModel(configuration) >>> model_pre = FlavaForPreTraining(configuration) >>> # Accessing the model configuration >>> configuration = model.config >>> configuration_pre = model_pre.config ``` r/Nr rT/L F@r?)r0rSrWr_rrprojection_dim init_codebooklogit_scale_init_valuerce_ignore_index mim_weight mlm_weightglobal_contrastive_weight itm_weightmmm_image_weightmmm_text_weightglobal_backprop_contrastive skip_unmasked_multimodal_encoder return_lossc! s|dd}|dd}|dd}|dd}tjf||dk r|dkrRi}tf|}|D]V\}}||krh|||krh|dkrh||krd|d|d}n d |d }t|qh|||dk r|dkri}t f|}d |krd d |d D|d <|D]`\}}||kr|||kr|dkr||krbd|d|d}n d|d }t|q|||dk r |dkri}t f|}|D]`\}}||kr|||kr|dkr||krd|d|d}n d|d }t|q|||dk r|dkr8i}t f|} | D]`\}}||krN|||krN|dkrN||krd|d|d}n d|d }t|qN|| |dkri}td|dkri}td|dkri}td|dkri}tdt f||_ tf||_ t f||_t f||_||_||_||_||_| |_| |_d|_| |_| |_| |_||_||_||_||_||_||_||_ dS)Ntext_config_dictimage_config_dictmultimodal_config_dictimage_codebook_config_dict)Ztransformers_version`zp` is found in both `text_config_dict` and `text_config` but with different values. The value `text_config_dict["z"]` will be used instead.zk`text_config_dict` is provided which will be used to initialize `FlavaTextConfig`. The value `text_config["z"]` will be overridden.Zid2labelcSsi|]\}}t||qSr))rF).0keyvaluer)r)r* `sz(FlavaConfig.__init__..zs` is found in both `image_config_dict` and `image_config` but with different values. The value `image_config_dict["zn`image_config_dict` is provided which will be used to initialize `FlavaImageConfig`. The value `image_config["z` is found in both `multimodal_config_dict` and `multimodal_config` but with different values. The value `multimodal_config_dict["z}`multimodal_config_dict` is provided which will be used to initialize `FlavaMultimodalConfig`. The value `multimodal_config["z` is found in both `image_codebook_config_dict` and `image_codebook_config` but with different values. The value `image_codebook_config_dict["z`image_codebook_config_dict` is provided which will be used to initialize `FlavaImageCodebookConfig`. The value `image_codebook_config["zR`image_config` is `None`. initializing the `FlavaImageConfig` with default values.zP`text_config` is `None`. Initializing the `FlavaTextConfig` with default values.z\`multimodal_config` is `None`. initializing the `FlavaMultimodalConfig` with default values.zc`image_codebook_config` is `None`. initializing the `FlavaImageCodebookConfig` with default values.rc)!popr#r$rJto_dictitemsr7infoupdaterrTrXr0rSrWr_rdrerrrrfZinitializer_factorrgrhrirjrkrlrmrnrorp)!r%r0rSrWr_rrrdrerfrrgrhrirjrkrlrmrnrorpr&rqrrrsrtZ_text_config_dictrwrxmessageZ_image_config_dictZ_multimodal_config_dictZ_image_codebook_config_dictr'r)r*r$s          "                          zFlavaConfig.__init__r0rSrWr_cKs(|f||||d|S)a& Instantiate a [`FlavaConfig`] (or a derived class) from flava text model configuration, flava image model configuration, flava multimodal model and flava codebook model configuration. Returns: [`FlavaConfig`]: An instance of a configuration object r)r{)r;r0rSrWr_r&r)r)r* from_configsszFlavaConfig.from_configs)NNNNr rr TrarrbrcrcrcrcrcrcTTT)r>r?r@rAr.rrFrrBrCrDr$rErrJrTrXrrIr)r)r'r*r`sfE    Kr`)rArGtypingrrrZconfiguration_utilsrutilsrZ get_loggerr>r7rrJrTrXr`r)r)r)r*s   wgU