fThSSKJr SSKJrJrJr SSKrSSKrSSKJ r SSK J r SSK J r SSKJrJr S S KJr S S KJr \R*"\5r\"S S\ 55r\"SS\ 55rS/rg)) dataclass)ListOptionalUnionN)Cache)$ImageClassifierOutputWithNoAttention)PreTrainedModel)auto_docstringlogging)AutoModelForImageTextToText)ShieldGemma2ConfigcB\rSrSr%SrSr\\R\ S'Sr g)0ShieldGemma2ImageClassifierOutputWithNoAttention$zVShieldGemma2 classifies imags as violative or not relative to a specific policy Args: N probabilities) __name__ __module__ __qualname____firstlineno____doc__rrtorchTensor__annotations____static_attributes__rn/var/www/auris/envauris/lib/python3.13/site-packages/transformers/models/shieldgemma2/modeling_shieldgemma2.pyrr$s-1M8ELL)0rrc ^\rSrSr\rS\4U4SjjrSrSrSr Sr Sr S r S r \SS \\R"S \\R$S \\R&S\\R"S\\\\R$\4S\\R"S\\R"S\\R$S\\R"S\\S\\S\\S\\S\\\R&4S\4Sjj5rSrU=r$)"ShieldGemma2ForImageClassification-configc>[TU]US9 [USS5Ul[USS5Ul[ R "US9Ulg)N)r$yes_token_indexi *no_token_indexi)super__init__getattrr&r'r from_configmodel)selfr$ __class__s r r)+ShieldGemma2ForImageClassification.__init__1sI '&v/@&I%f.>E0<>@@rcNURRRU5 gr1)r,r2set_output_embeddings)r-new_embeddingss r r>8ShieldGemma2ForImageClassification.set_output_embeddings@s !!77GrcNURRRU5 gr1)r,r2 set_decoder)r-decoders r rB.ShieldGemma2ForImageClassification.set_decoderCs !!--g6rcJURRR5$r1)r,r2 get_decoderr4s r rF.ShieldGemma2ForImageClassification.get_decoderFzz((4466rcJURRR5$r1)r,r2 tie_weightsr4s r rJ.ShieldGemma2ForImageClassification.tie_weightsIrHr input_ids pixel_valuesattention_mask position_idspast_key_valuestoken_type_idscache_position inputs_embedslabels use_cacheoutput_attentionsoutput_hidden_states return_dictlogits_to_keepreturnc UR"SUUUUUUUUU U U U U US.UD6nURnUSS2SURUR/4n[R "USS9n[ UUS9$)a Returns: A `ShieldGemma2ImageClassifierOutputWithNoAttention` instance containing the logits and probabilities associated with the model predicting the `Yes` or `No` token as the response to that prompt, captured in the following properties. * `logits` (`torch.Tensor` of shape `(batch_size, 2)`): The first position along dim=1 is the logits for the `Yes` token and the second position along dim=1 is the logits for the `No` token. * `probabilities` (`torch.Tensor` of shape `(batch_size, 2)`): The first position along dim=1 is the probability of predicting the `Yes` token and the second position along dim=1 is the probability of predicting the `No` token. ShieldGemma prompts are constructed such that predicting the `Yes` token means the content *does violate* the policy as described. If you are only interested in the violative condition, use `violated = outputs.probabilities[:, 1]` to extract that slice from the output tensors. When used with the `ShieldGemma2Processor`, the `batch_size` will be equal to `len(images) * len(policies)`, and the order within the batch will be img1_policy1, ... img1_policyN, ... imgM_policyN. )rLrMrNrOrPrQrRrSrTrUrVrWrXrYN)dim)logitsrr)r,r^r&r'rsoftmaxr)r-rLrMrNrOrPrQrRrSrTrUrVrWrXrY lm_kwargsoutputsr^selected_logitsrs r forward*ShieldGemma2ForImageClassification.forwardLsN** %)%+))'/!5#)  " B)=)=t?R?R(S!ST o2> ?"'  r)r,r'r&)NNNNNNNNNNNNNr)rrrrr config_classr)r3r7r;r>rBrFrJr rr LongTensor FloatTensorrrrrboolintrrcr __classcell__)r.s@r r"r"-s%LL1L @>AH77715481537KO595959-1$(,0/3&*34= E,,-= u001= !. = u//0 = "%U->->(?(F"GH = !!1!12= !!1!12=   1 12= ))*= D>= $D>= 'tn= d^= c5<</0= " :#= = rr") dataclassesrtypingrrrrtorch.utils.checkpoint cache_utilsrmodeling_outputsr modeling_utilsr utilsr r autorconfiguration_shieldgemma2r get_loggerrloggerrr"__all__rrr rws "((  D-/:   H % 17[1 1\ \ \ @) r