o ei=@sLddlmZddlZddlmZddlmZddlmZm Z ddl m Z m Z ddl mZmZddlmZdd lmZdd lmZdd lmZd d lmZd dlmZmZmZmZmZm Z m!Z!m"Z"Gddde Z#Gddde Z$GdddeZ%GdddeZ&GdddeZ'GdddeZ(GdddeZ)GdddeZ*gdZ+dS) )CallableN)TransformersKwargs)Cache DynamicCache)PreTrainedConfiglayer_type_validation)create_causal_mask!create_sliding_window_causal_mask)BaseModelOutputWithPast)RopeParameters)ALL_ATTENTION_FUNCTIONS)Unpack)Gemma2RotaryEmbedding)Olmo2AttentionOlmo2DecoderLayerOlmo2ForCausalLM Olmo2ModelOlmo2PreTrainedModel Olmo2RMSNormapply_rotary_pos_embeager_attention_forwardc*s@eZdZdZdZdgZddddddddZd gd gfd d gd gfd gd gfd Z                    d2dedBdedBdedBdedBd edBd!edBd"e dBd#edBd$e dBd%e dBd&edBd'edBd(edBd)e dBd*e e e e fBdBd+e dBd,e dBd-e dBd.edBd/ee dBf(fd0d1 ZZS)3 Olmo3Configa This is the configuration class to store the configuration of a [`Olmo3Model`]. It is used to instantiate an OLMo3 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 [allenai/OLMo-3-0725-1B](https://huggingface.co/allenai/OLMo-3-0725-1B). 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 50304): Vocabulary size of the Olmo3 model. Defines the number of different tokens that can be represented by the `inputs_ids` passed when calling [`Olmo3Model`] hidden_size (`int`, *optional*, defaults to 4096): Dimension of the hidden representations. intermediate_size (`int`, *optional*, defaults to 11008): Dimension of the MLP representations. num_hidden_layers (`int`, *optional*, defaults to 32): Number of hidden layers in the Transformer decoder. num_attention_heads (`int`, *optional*, defaults to 32): Number of attention heads for each attention layer in the Transformer decoder. num_key_value_heads (`int`, *optional*): This is the number of key_value heads that should be used to implement Grouped Query Attention. If `num_key_value_heads=num_attention_heads`, the model will use Multi Head Attention (MHA), if `num_key_value_heads=1` the model will use Multi Query Attention (MQA) otherwise GQA is used. When converting a multi-head checkpoint to a GQA checkpoint, each group key and value head should be constructed by meanpooling all the original heads within that group. For more details, check out [this paper](https://huggingface.co/papers/2305.13245). If it is not specified, will default to `num_attention_heads`. hidden_act (`str` or `function`, *optional*, defaults to `"silu"`): The non-linear activation function (function or string) in the decoder. max_position_embeddings (`int`, *optional*, defaults to 2048): The maximum sequence length that this model might ever be used with. initializer_range (`float`, *optional*, defaults to 0.02): The standard deviation of the truncated_normal_initializer for initializing all weight matrices. use_cache (`bool`, *optional*, defaults to `True`): Whether or not the model should return the last key/values attentions (not used by all models). Only relevant if `config.is_decoder=True`. pad_token_id (`int`, *optional*, defaults to 1): Padding token id. bos_token_id (`int`, *optional*): Beginning of stream token id. eos_token_id (`int`, *optional*, defaults to 50279): End of stream token id. tie_word_embeddings (`bool`, *optional*, defaults to `False`): Whether to tie weight embeddings rope_parameters (`RopeParameters`, *optional*): Dictionary containing the configuration parameters for the RoPE embeddings. The dictionary should contain a value for `rope_theta` and optionally parameters used for scaling in case you want to use RoPE with longer `max_position_embeddings`. attention_bias (`bool`, defaults to `False`, *optional*, defaults to `False`): Whether to use a bias in the query, key, value and output projection layers during self-attention. attention_dropout (`float`, *optional*, defaults to 0.0): The dropout ratio for the attention probabilities. rms_norm_eps (`float`, *optional*, defaults to 1e-05): The epsilon used by the rms normalization layers. sliding_window (`int`, *optional*, defaults to 4096): Size of the sliding window for sliding window attention. layer_types (`list`, *optional*): Attention pattern for each layer. Defaults to sliding window attention for 3 out of 4 layers, and full attention for every 4th layer. ```python >>> from transformers import Olmo3Model, Olmo3Config >>> # Initializing a Olmo3 7B style configuration >>> configuration = Olmo3Config() >>> # Initializing a model from the Olmo3 7B style configuration >>> model = Olmo3Model(configuration) >>> # Accessing the model configuration >>> configuration = model.config ``` olmo3past_key_valuescolwise_gather_outputrowwise_split_inputcolwiserowwise)zlayers.*.self_attn.q_projzlayers.*.self_attn.k_projzlayers.*.self_attn.v_projzlayers.*.self_attn.o_projzlayers.*.mlp.gate_projzlayers.*.mlp.up_projzlayers.*.mlp.down_proj input_ids inputs_embeds hidden_statesattention_mask) embed_tokenslayersnorm+ Nsilu{Gz?TgFh㈵> vocab_size hidden_sizeintermediate_sizenum_hidden_layersnum_attention_headsnum_key_value_heads hidden_actmax_position_embeddingsinitializer_range use_cache pad_token_id bos_token_id eos_token_idtie_word_embeddingsrope_parametersattention_biasattention_dropout rms_norm_epssliding_window layer_typesc s||_||_||_||_||_||_|dur|}||_||_| |_| |_ ||_ ||_ ||_ | |_ | |_| |_||_||_||_|jdurOddt|jD|_t|j|j||_tjdi|dS)NcSs$g|]}|dddkrdndqS)r.rsliding_attentionfull_attention).0irIrIe/home/ubuntu/transcripts/venv/lib/python3.10/site-packages/transformers/models/olmo3/modular_olmo3.py sz(Olmo3Config.__init__..rI)r2r9r3r4r5r6r7r8r:r;rArBr?r<r=r>rCrDrErangerr@super__init__)selfr2r3r4r5r6r7r8r9r:r;r<r=r>r?r@rArBrCrDrEkwargs __class__rIrLrPs8 zOlmo3Config.__init__)r'r(r)r*r*Nr+r,r-Tr.Nr/FNFr0r1r(N)__name__ __module__ __qualname____doc__ model_typekeys_to_ignore_at_inferencebase_model_tp_planbase_model_pp_planintstrfloatboolr dictlistrP __classcell__rIrIrSrLr*sL        rc@ eZdZdS) Olmo3RMSNormNrUrVrWrIrIrIrLrerecseZdZdedeffdd Z  ddejdeejejfdejdBd e dBd ej dBd e e d eejejdBffd dZ ZS)Olmo3Attentionconfig layer_idxcsJtj||d|jdusJ|j||_|jdkr |j|_dSd|_dS)N)rjrG)rOrPrEattention_typerD)rQrirjrSrIrLrPs  zOlmo3Attention.__init__Nr"position_embeddingsr#rcache_positionrRreturncKs4|jdd}g|d|jR}|||} |||} ||} | |dd} | |dd} | |dd} |\} } t | | | | \} } |durc| | |d}| | | |j |\} } t |jjt}||| | | |f|jswdn|j|j|jd|\}}|jg|dR}||}||fS)Nr.r)sincosrmr0)dropoutscalingrD)shapehead_dimq_normq_projk_normk_projv_projview transposerupdaterjr get_interfaceri_attn_implementationrtrainingrBrsrDreshape contiguouso_proj)rQr"rlr#rrmrR input_shape hidden_shape query_states key_states value_statesrqrp cache_kwargsattention_interface attn_output attn_weightsrIrIrLforwards@     zOlmo3Attention.forward)NN)rUrVrWrr]rPtorchTensortupler LongTensorrrrrcrIrIrSrLrhs& rhc@rd)Olmo3DecoderLayerNrfrIrIrIrLrrgrc@rd)Olmo3RotaryEmbeddingNrfrIrIrIrLrrgrc@rd)Olmo3PreTrainedModelNrfrIrIrIrLr rgrcseZdZdeffdd Z       ddejdBdejdBdejdBdedBd ej dBd ejdBd e dBd e e d e fddZZS) Olmo3ModelricsNttjjd|_tfddtj D|_ t d|_ dS)N)epscsg|]}t|qSrI)r)rJrjrirIrLrMsz'Olmo3Model.__init__..r) rOrPrer3rCr&nn ModuleListrNr5r%r rotary_emb)rQrirSrrLrPs zOlmo3Model.__init__Nr r# position_idsrr!rmr;rRrnc Ks,|du|duAr td|dur||}|r!|dur!t|jd}|dur<|dur-|nd} tj|jd|jd| }|durE| d}t |} t se|j|||||d} t d i| t d i| d} |} || |} |jd|jjD]}|| f| |jj|||| d|} qv|| } t| |d S) Nz:You must specify exactly one of input_ids or inputs_embedsrrr.)device)rir!r#rmrr)rHrG)r#rrrmrl)last_hidden_staterrI) ValueErrorr$rriget_seq_lengthrarangertr unsqueeze isinstancerar r rr%r5 self_attnrkr&r )rQr r#rrr!rmr;rRpast_seen_tokenscausal_mask_mapping mask_kwargsr"rl decoder_layerrIrIrLrsT         zOlmo3Model.forward)NNNNNNN)rUrVrWrrPrrrr FloatTensorr`rrr rrcrIrIrSrLrs8   rc@rd)Olmo3ForCausalLMNrfrIrIrIrLr]rgr)rrrr),collections.abcrrtorch.nnrtransformers.utils.genericr cache_utilsrrconfiguration_utilsrr masking_utilsr r modeling_outputsr modeling_rope_utilsr modeling_utilsr processing_utilsrgemma2.modeling_gemma2rolmo2.modeling_olmo2rrrrrrrrrrerhrrrrr__all__rIrIrIrLs.        ( 8L