^j{IdZddlmZddlmZddlmZddlmZm Z e je Z edeGd d eZ edeGd d eZedeGd deZedeGddeZedeGddeZgdZy)zFLAVA model configurations)Any)strict)PreTrainedConfig)auto_docstringloggingzfacebook/flava-full) checkpointc0eZdZUdZdZdZdZeed<dZ eed<dZ eed<d Z eed <d Z e ed <d Zeezed<d Zeezed<dZeed<dZeed<dZeeezeeefzed<dZeeezeeefzed<dZeed<dZeed<dZeed<dZeed<y) FlavaImageConfigaT mask_token (`bool`, *optional*, defaults to `True`): Whether to use a mask token or not. Used in 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 ```flava_image_model image_config hidden_size num_hidden_layersnum_attention_heads intermediate_sizegelu hidden_acthidden_dropout_probattention_probs_dropout_prob{Gz?initializer_range-q=layer_norm_eps image_size patch_sizer num_channelsTqkv_bias mask_token vocab_sizeN)__name__ __module__ __qualname____doc__ model_typebase_config_keyrint__annotations__rrrrstrrfloatrrrrlisttupler!r"r#boolr$r&x/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/transformers/models/flava/configuration_flava.pyr r s &%J$OKs!!!s!J'**03 %#+3#u#!NE!47Jd3i%S/1746Jd3i%S/16L#HdJJr5r ceZdZUdZdZdZdZeed<dZ eed<dZ eed <d Z eed <d Z eed <d Z eed<dZeed<dZeed<dZeezed<dZeezed<dZeed<dZeed<dZedzed<dZeed<y)FlavaTextConfiga 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 ```flava_text_model text_configi:wr&type_vocab_sizeimax_position_embeddingsrrrrrrrrrrrrrrrrrN pad_token_idTr#)r'r(r)r*r+r,r&r-r.r<r=rrrrrr/rr0rrrr>r#r3r4r5r6r8r8Es $J#OJOS#&S&Ks!!!s!J'**03 %#+3#u#!NE! L#* Hdr5r8ceZdZUdZdZdZdZeed<dZ eed<dZ eed <d Z eed <d Z e ed <dZeezed<dZeezed<dZeed<dZeed<dZeed<dZeed<y)FlavaMultimodalConfiga{ 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 ```flava_multimodal_modelmultimodal_configrrrrrrrrrrrrrrrrTr# use_cls_tokenN)r'r(r)r*r+r,rr-r.rrrrr/rr0rrrr#r3rDr4r5r6r@r@ks &*J)OKs!!!s!J'**03 %#+3#u#!NE!HdM4r5r@cteZdZUdZdZeed<dZeed<dZeed<dZ eed <d Z eed <d Z e ed <dZ eed<y)FlavaImageCodebookConfigas num_groups (`int`, *optional*, defaults to 4): Number of groups to be created. This parameter as of now doesn't affect the model and is used for some internal calculation and estimations. num_blocks_per_group (`int`, *optional*, defaults to 2): Number of conv-based blocks per group. freeze (`bool`, defaults to `True`): Whether to freeze the weights of the model. Example: ```python >>> from transformers import FlavaImageCodebookConfig, FlavaImageCodebook >>> # Initializing a FlavaImageCodebook with style configuration >>> configuration = FlavaImageCodebookConfig() >>> # Initializing a FlavaImageCodebook model (with random weights) from the style configuration >>> model = FlavaImageCodebook(configuration) >>> # Accessing the model configuration >>> configuration = model.config ```  num_groupsrinput_channelsr;num_blocks_per_grouprr%r&TfreezerrN)r'r(r)r*rHr-r.rIrJrr&rLr3rr0r4r5r6rFrFsQ0JNC !#!KJFD#u#r5rFceZdZUdZdZeeeedZ dZ e e e fezdzed<dZe e e fezdzed<dZe e e fezdzed<dZe e e fezdzed<d Zeed <d Zeed <d Zeed <dZeed<dZeed<dZeed<dZeed<dZeed<dZeed<dZeed<dZ eed<dZ!eed<dZ"eed<dZ#eed<dZ$eed<dZ%eed<dZ&eed <dZ'eed!<fd"Z(xZ)S)# FlavaConfiga image_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`FlavaImageConfig`]. multimodal_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`FlavaMultimodalConfig`]. image_codebook_config (`dict`, *optional*): Dictionary of configuration options used to initialize [`FlavaCodebookConfig`]. init_codebook (`bool`, *optional*, defaults to `True`): Whether to initialize the codebook 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. 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 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 ``` flava)r:r rBimage_codebook_configNr r:rBrPrrrrprojection_dimT init_codebookg/L F@logit_scale_init_valuerrice_ignore_indexg? mim_weight mlm_weightglobal_contrastive_weight itm_weightmmm_image_weightmmm_text_weightglobal_backprop_contrastive skip_unmasked_multimodal_encoder return_losstie_word_embeddingsinitializer_factorc |ji}tjdnAt|jtr|jj }n |j}|j i}tjdnAt|j tr|j j }n |j }|ji}tjdnAt|jtr|jj }n |j}|ji}tjdnAt|jtr|jj }n |j}|jdd}|jdd}|jdd}|jdd} |t di|j } | jD]B\} } | |vs | || k7s| d k7s| |vr d | d | d } nd | d} tj| D|j| |tdi|j }d|vr3|djD cic]\} } t| | c} } |d<|jD]B\} } | |vs | || k7s| d k7s| |vr d | d| d } nd| d} tj| D|j||tdi|j }|jD]B\} } | |vs | || k7s| d k7s| |vr d | d| d } nd| d} tj| D|j|| tdi| j }|jD]B\} } | |vs | || k7s| d k7s| | vr d | d| d } nd| d} tj| D|j|t di||_tdi||_tdi||_tdi||_ t!|Ddi|ycc} } w)NzP`text_config` is `None`. Initializing the `FlavaTextConfig` with default values.zR`image_config` is `None`. initializing the `FlavaImageConfig` 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.text_config_dictimage_config_dictmultimodal_config_dictimage_codebook_config_dicttransformers_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.id2labelzs` 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["r4)r:loggerinfo isinstancer8to_dictr r rBr@rPrFpopitemsupdater/super __post_init__)selfkwargsr:r rBrPrarbrcrd_text_config_dictkeyvaluemessage_image_config_dict_multimodal_config_dict_image_codebook_config_dict __class__s r6rpzFlavaConfig.__post_init__ s    #K KKj k ((/ :**224K**K    $L KKl m ))+; <,,446L,,L  ! ! ) "  KKv w ..0E F $ 6 6 > > @  $ 6 6   % % -$& ! KKu 224L M$($>$>$F$F$H !$($>$> !"::&8$?"JJ':DA!',Dd!K%+ZZ0Ld%S"  ' / C2B C K K M 0557 ) U+%%;s3C*COeHe..u%<T&U&]&]&_ #6;;= ) U++9J39O0OTW[qTq44u%TTWSXXqs VVYUZZqs KK( )"  $ $%< = % 1*B*`E_*`*h*h*j ':??A ) U00!6s!;;5588u%]]`\ab'' ]]`\aaxz KK(' ), " ( ()D E+9[9,<|<!6!K9J!K%=%V@U%V" ''k2sQ))*r'r(r)r*r+r8r r@rF sub_configsr dictr/rrr.r:rBrPrr-rr0rQrRr3rSrrTrUrVrWrXrYrZr[r\r]r^r_rp __classcell__)rzs@r6rNrNsd1fJ&(2!9 K>BL$sCx.#33d:A<@Kc3h"22T9@BFtCH~(884?FFJ4S>,<rs !.3,   H %01%'%2%P01!&!2!H01!,!2!H01$/$2$D01q("q(2q(h vr5