`iLddlmZmZddlZddlmZddlmZddlm Z m Z ddl m Z ddl mZmZddlmZdd lmZmZdd lmZdd lmZd d lmZd dlmZmZmZmZm Z m!Z!m"Z"m#Z#m$Z$GddeZ%Gdde!Z&GddeZ'GddeZ(Gdde"Z)Gdde Z*GddeZ+GddeZ,gdZ-dS) )CallableOptionalN)TransformersKwargs)Cache DynamicCache)layer_type_validation)create_causal_mask!create_sliding_window_causal_mask)BaseModelOutputWithPast)ROPE_INIT_FUNCTIONSrope_config_validation)ALL_ATTENTION_FUNCTIONS)Unpack) Olmo2Config) Olmo2AttentionOlmo2DecoderLayerOlmo2ForCausalLM Olmo2ModelOlmo2PreTrainedModel Olmo2RMSNormOlmo2RotaryEmbeddingapply_rotary_pos_embeager_attention_forwardceZdZdZdZddddddddZdgd gfd d gd gfd gd gfd Z dfd ZdZxZ S) 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_theta (`float`, *optional*, defaults to 10000.0): The base period of the RoPE embeddings. rope_scaling (`Dict`, *optional*): Dictionary containing the scaling configuration for the RoPE embeddings. NOTE: if you apply new rope type and you expect the model to work on longer `max_position_embeddings`, we recommend you to update this value accordingly. Expected contents: `rope_type` (`str`): The sub-variant of RoPE to use. Can be one of ['default', 'linear', 'dynamic', 'yarn', 'longrope', 'llama3'], with 'default' being the original RoPE implementation. `factor` (`float`, *optional*): Used with all rope types except 'default'. The scaling factor to apply to the RoPE embeddings. In most scaling types, a `factor` of x will enable the model to handle sequences of length x * original maximum pre-trained length. `original_max_position_embeddings` (`int`, *optional*): Used with 'dynamic', 'longrope' and 'llama3'. The original max position embeddings used during pretraining. `attention_factor` (`float`, *optional*): Used with 'yarn' and 'longrope'. The scaling factor to be applied on the attention computation. If unspecified, it defaults to value recommended by the implementation, using the `factor` field to infer the suggested value. `beta_fast` (`float`, *optional*): Only used with 'yarn'. Parameter to set the boundary for extrapolation (only) in the linear ramp function. If unspecified, it defaults to 32. `beta_slow` (`float`, *optional*): Only used with 'yarn'. Parameter to set the boundary for interpolation (only) in the linear ramp function. If unspecified, it defaults to 1. `short_factor` (`list[float]`, *optional*): Only used with 'longrope'. The scaling factor to be applied to short contexts (< `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden size divided by the number of attention heads divided by 2 `long_factor` (`list[float]`, *optional*): Only used with 'longrope'. The scaling factor to be applied to long contexts (< `original_max_position_embeddings`). Must be a list of numbers with the same length as the hidden size divided by the number of attention heads divided by 2 `low_freq_factor` (`float`, *optional*): Only used with 'llama3'. Scaling factor applied to low frequency components of the RoPE `high_freq_factor` (`float`, *optional*): Only used with 'llama3'. Scaling factor applied to high frequency components of the RoPE 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 ``` olmo3 colwise_rep rowwise_repcolwiserowwise)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?TgF@h㈵>c Dtjdid|d|d|d|d|d|d|d|d | d | d | d | d | d|d|d|d|d|d||||_||_|j#dt |jD|_t |jdS)N 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_embeddings rope_theta rope_scalingattention_biasattention_dropout rms_norm_epsc.g|]}|dzdzdkrdndS)r1rsliding_attentionfull_attention).0is {/home/jaya/work/projects/VOICE-AGENT/VIET/agent-env/lib/python3.11/site-packages/transformers/models/olmo3/modular_olmo3.py z(Olmo3Config.__init__..s>   RSA{a'7'7##=M   rN)super__init__sliding_window layer_typesranger:r )selfr7r8r9r:r;r<r=r>r?r@rArBrCrDrErFrGrHrIrVrWkwargs __class__s rQrUzOlmo3Config.__init__sv2    !z #  0/ 0/  !4 3  !4 3  "z %<$; 0/  i & & & !4 3 "z &! "*># $0/% &&)   .-&   #  W\]a]sWtWt   D  d./////rSc$t|dS)z< Validate the `rope_scaling` configuration. N)r)rYs rQ_rope_scaling_validationz$Olmo3Config._rope_scaling_validations t$$$$$rS)r*r+r,r-r-Nr.r/r0Tr1Nr2Fr3NFr4r5r+N) __name__ __module__ __qualname____doc__ model_typebase_model_tp_planbase_model_pp_planrUr] __classcell__r[s@rQrr,smm^J%2%2%2%2"+ )"+&(9:#%568IJ!"_$56  $!-606060606060p%%%%%%%rSrceZdZdS) Olmo3RMSNormNr^r_r`rNrSrQrhrhDrSrhceZdZdedeffd Z d dejdeejejfde ejde e d e ej d e e d eeje ejffd ZxZS)Olmo3Attentionconfig layer_idxct|||jJ|j||_|jdkr|jnd|_dS)N)rnrL)rTrUrWattention_typerV)rYrmrnr[s rQrUzOlmo3Attention.__init__sd 9555!---$0;7;7JNa7a7af33gkrSNr%position_embeddingsr&past_key_valuescache_positionrZreturnc |jdd}g|d|jR}|||} |||} ||} | |dd} | |dd} | |dd} |\} } t| | | | \} } |&| | |d}| | | |j |\} } t}|j jdkrt|j j}||| | | |f|jsdn|j|j|jd|\}}|jg|dR}||}||fS)Nr1r)sincosrseagerr4)dropoutscalingrV)shapehead_dimq_normq_projk_normk_projv_projview transposerupdaternrrm_attn_implementationrtrainingrHr{rVreshape contiguouso_proj)rYr%rqr&rrrsrZ input_shape hidden_shape query_states key_states value_statesrxrw cache_kwargsattention_interface attn_output attn_weightss rQforwardzOlmo3Attention.forwards$)#2#. 88b8$-88 {{4;;}#=#=>> [[]!;!;<< {{=11 #((66@@AFF __\22<.)!. .) "% .) !!12 .)+,.) u|Xel33 4.).).).).).).).)rSrlceZdZdS)Olmo3DecoderLayerNrirNrSrQrr)rjrSrc,eZdZddedeefdZdS)Olmo3RotaryEmbeddingNrm rope_typec8tj||||_njt |drSt |jtr9|jd|jd|_nd|_|jJ|j |_ |j |_ ||_ t|j|_||j |\}|_|d|d|j|_dS)NrFrtypedefaultinv_freqF) persistent)nnModulerUrhasattr isinstancerFdictgetr>max_seq_len_cachedoriginal_max_seq_lenrmr rope_init_fnattention_scalingregister_bufferroriginal_inv_freq)rYrmdevicerrs rQrUzOlmo3RotaryEmbedding.__init__0s 4    &DNN V^ , , 'F.Rs$ c c ci vy 1 1 c c crSr)rmrrmrLrM)rTrUrhr8rIr)r ModuleListrXr:r( ModuleDictr rotary_embs rotary_emb)rYrmr[s `rQrUzOlmo3Model.__init__Ns     !39LMMM m c c c c5IaCbCb c c c   =%9S\%]%]%]"6f"E"E"E     OOOrSNr#r& position_idsrrr$rsr@rZrtc |du|duz rtd|||}|r|t|j}|B||nd} t j| | |jdz|j}|| d}t|x} ts'|j|||||d} td i| td i| d} |} |jd| ||jd | |d } |jd|jjD]1}|| f| |jj|||| |jjd |} 2|| } t)| | S)Nz:You must specify exactly one of input_ids or inputs_embedsrrr1)r)rm input_embedsr&rsrrr)rMrLrLrMr)r&rrrrsrq)last_hidden_staterrrN) ValueErrorr'rrmget_seq_lengthraranger|r unsqueezerrr r rr(r: self_attnrpr)r )rYr#r&rrrr$rsr@rZpast_seen_tokenscausal_mask_mapping mask_kwargsr%position_embeddings_mapping decoder_layers rQrzOlmo3Model.forward\s%  -t"; < [YZZ Z  *.*;*;I*F*FM  ?0*$+>>>O  !CRC^==???de +0< "2]5H5K"KTaTh,,,N  )33A66L?-FF + -"0"0#2 , K#5"C"C{"C"C%F%U%U%U%U##  & !F!12E!F}Vb!c!c@d./?@P\]]' ' # "[)H4;+H)HI  M)M2=3J3YZ) /-$? @W@f$g MM -00 &++    rS)NNNNNNN)r^r_r`rrUrrrrr FloatTensorboolrrr rrerfs@rQrrMs {       151537+/5959$(C C E,-C !.C u/0 C "% C   12 C !!12C D>C +,C  !C C C C C C C C rSrceZdZdS)Olmo3ForCausalLMNrirNrSrQrrrjrSr)rrrr).typingrrrtorch.nnrtransformers.utils.genericr cache_utilsrrconfiguration_utilsr masking_utilsr r modeling_outputsr modeling_rope_utilsr rmodeling_utilsrprocessing_utilsrolmo2.configuration_olmo2rolmo2.modeling_olmo2rrrrrrrrrrrhrlrrrrr__all__rNrSrQrs &%%%%%%% 999999........888888RRRRRRRR777777NNNNNNNN555555&&&&&&333333                      |%|%|%|%|%+|%|%|%~     <   5)5)5)5)5)^5)5)5)p     )   /////////.     /   R R R R R R R R j     '      rS