h-] B d Z ddlmZmZmZ ddlZddlmZ ddlmZ ddl m
Z
m Z ddl m
Z
mZ dd lmZmZ dd
lmZmZ dd lmZ dd lmZ dd
lmZmZ ddl m Z ddl m Z m Z m!Z!m"Z"m#Z#m$Z$m%Z%m&Z&m'Z'm(Z( ddl)m*Z*m+Z+ ejX e- Z.dZ/dZ0 G d de
Z1 G d de% Z2 G d de& Z3 G d dejh Z5 G d d e+ Z6 G d d e* Z7 G d d e$ Z8 G d! d"e8e# Z9 G d# d$e Z: G d% d&e! Z; G d' d(e" Z< G d) d*e Z=g d+Z>y),zLG AI Research EXAONE Lab )CallableOptionalUnionN)nn)check_model_inputs )Cache DynamicCache)PretrainedConfiglayer_type_validation)create_causal_mask!create_sliding_window_causal_mask)BaseModelOutputWithPastCausalLMOutputWithPast)ALL_ATTENTION_FUNCTIONS)Unpack)TransformersKwargslogging)deprecate_kwarg )
LlamaForCausalLMLlamaForQuestionAnswering LlamaForSequenceClassificationLlamaForTokenClassification
LlamaModelLlamaPreTrainedModel LlamaRMSNormLlamaRotaryEmbeddingapply_rotary_pos_embeager_attention_forward)Olmo2DecoderLayerOlmo2MLPz LGAI-EXAONE/EXAONE-4.0-Instruct
Exaone4Configc e Zd ZdZdZdgZddddddddZdgdgfd d
gd gfd gd gfd Z d
fd Z xZ S )r# a
This is the configuration class to store the configuration of a [`Exaone4Model`]. It is used to
instantiate a EXAONE 4.0 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 EXAONE-4.0-Instruct [LGAI-EXAONE/EXAONE-4.0-Instruct](https://huggingface.co/LGAI-EXAONE/EXAONE-4.0-Instruct)
NOTE: `EXAONE-4.0-Instruct` is a placeholder model ID. The exact model ID will be updated in the future.
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 102400):
Vocabulary size of the EXAONE 4.0 model. Defines the number of different tokens that can be represented by the
`inputs_ids` passed when calling [`Exaone4Model`].
hidden_size (`int`, *optional*, defaults to 4096):
Dimension of the hidden representations.
intermediate_size (`int`, *optional*, defaults to `hidden_size * 4`):
Dimensionality of the MLP representations.
num_hidden_layers (`int`, *optional*, defaults to 32):
Number of hidden layers in the Transformer encoder.
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 checkout [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. Typically set this to something large
just in case (e.g., 32768 for EXAONE 3.5).
initializer_range (`float`, *optional*, defaults to 0.02):
The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
rms_norm_eps (`float`, *optional*, defaults to 1e-05):
The epsilon used by the layer normalization layers.
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``.
bos_token_id (`int`, *optional*, defaults to 0):
Beginning of stream token id.
eos_token_id (`int`, *optional*, defaults to 2):
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_dropout (`float`, *optional*, defaults to 0.0):
The dropout ratio for the attention probabilities.
sliding_window (`int`, *optional*):
The size of the sliding window for the sliding window attention.
sliding_window_pattern (`str`, *optional*):
The pattern to use for sliding window attention. Can be one of:
- `None`: No sliding window attention is used
- `int`: Every `sliding_window` layers, use global attention, else use local attention.
- `str`: A sequence of "L" (local attention) and "G" (global attention) characters that defines the
attention pattern. The pattern starts from layer 0 and repeats every `sliding_window` layers. The
final layer always uses global attention regardless of the pattern.
For instance, sliding_window_pattern="LLLG" same as sliding_window=4, which means:
- Layer 0, 1, 2: local attention,
- Layer 3: global attention,
...(repeated)
layer_types (`list`, *optional*):
Attention pattern for each layer. Prioritized over `sliding_window_pattern`.
Example:
```python
>>> from transformers import Exaone4Model, Exaone4Config
>>> # Initializing a EXAONE configuration
>>> configuration = Exaone4Config()
>>> # Initializing a model from configuration
>>> model = Exaone4Model(configuration)
>>> # Accessing the model configuration
>>> configuration = model.config
```exaone4past_key_valuescolwiserowwise)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_tokenslayersnormc 4 || _ || _ || _ || _ || _ || _ || _ || _ | | _ |
| _ | | _
|| _ || _ || _
|| _ || _ || _ | j d}| j Dt# | j D cg c] }|dz |z dk7 r|| j k rdnd" c}| _ d| j v rd| _ t' | j t) | T d| |
|d| y c c}w ) Nr sliding_attentionfull_attentionsliding_windowhybrid) bos_token_id eos_token_idtie_word_embeddings )
vocab_size hidden_sizenum_hidden_layersnum_attention_headsnum_key_value_headsintermediate_size
hidden_actmax_position_embeddingsinitializer_range rms_norm_eps use_cacheattention_dropout
rope_theta rope_scalingr4 sliding_window_pattern layer_typesrangecache_implementationr super__init__)selfr: r; r? r< r= r> r@ rA rB rC rD r6 r7 r8 rF rG rE r4 rH rI kwargsi __class__s i/var/www/html/eduruby.in/venv/lib/python3.12/site-packages/transformers/models/exaone4/modular_exaone4.pyrM zExaone4Config.__init__ sG 0 %&!2 #6 #6 !2 $'>$!2 ("!2 $ ( ,&<#& &%& " #
t 5 56 U 56!;DDZDZ@Z $%& D t / / /(0D % d . ./
%LVi
ms
s *%D)i i @ rT rT silui g{Gz?gh㈵>Tr r Fg @N rS N)
__name__
__module__ __qualname____doc__
model_typekeys_to_ignore_at_inferencebase_model_tp_planbase_model_pp_planrM
__classcell__rQ s @rR r# r# < s u n J#4"5 &/%.%.%."+ )"+ & (9:#%56 8IJ!"_$56 $ ! +9
9
c e Zd Zy)Exaone4RMSNormNrX rY rZ r9 rb rR rd rd rb rd c e Zd Zy)Exaone4RotaryEmbeddingNre r9 rb rR rh rh rf rb rh c P e Zd Zdedef fd Z eddd ddej d e ej ej f d
e
ej de
e d e
ej d e
e d
e ej e
ej e
e ej f fd Z xZS )Exaone4Attentionconfig layer_idxc t | || _ || _ |j | _ |j
| _ |j | _ t |d|j |j z | _ |j |j
z | _ |j | _
d| _ | j dz | _ |j | _
|j | _ |j | dk( | _ t# j$ | j | j | j z d | _ t# j$ | j | j
| j z d | _ t# j$ | j | j
| j z d | _ t# j$ | j | j z | j d | _ t/ | j |j0 | _ t/ | j |j0 | _ y )Nhead_dimTg r2 F)biaseps)rL rM rk rl r= r> r; getattrrn num_key_value_groupsrE is_causalscalingr4 rH rI
is_slidingr Linearq_projk_projv_projo_projrd rC q_normk_normrN rk rl rQ s rR rM zExaone4Attention.__init__ s
"#)#=#= #)#=#= !--
F4F4F&JdJd4de
$*$>$>&B\B\$\!!'!9!9 }}d* $ 3 3&,&C&C# ,,Y7;NNii 0 0$2J2JT]]2Zafg ii 0 0$2J2JT]]2Zafg ii 0 0$2J2JT]]2Zafg ii 8 84== H$JZJZafg $T]]8K8KL $T]]8K8KL rb past_key_valuer& z4.58)new_nameversionr+ position_embeddingsr, cache_positionrO returnc |j d d }g |d| j }| j | j | j dd } | j | j | j dd }
| j
| j | j dd } | j | } | j |
}
|\ } }
| j | j rt | |
| |
\ } }
|%d|i}|j |
| | j | \ }
} t }| j j dk7 r t" | j j } || | |
| |f| j$ sdn | j& | j( | j r | j nd d|\ }} |j* g |d j- }| j/ | }||fS )Nr1 r r eagerrV )dropoutru r4 )shapern rx view transposery rz r| r} r4 rv r updaterl r rk _attn_implementationr trainingrE ru reshape
contiguousr{ )rN r+ r r, r& r rO input_shape hidden_shape query_states
key_states value_statescossin cache_kwargsattention_interface attn_output attn_weightss rR forwardzExaone4Attention.forward# s $))#2. 88b8$--8 {{=166|DNNqRST [[ /44\BLLQPQR
{{=166|DNNqRST {{<0 [[,
&S &$//';L*VY[^'_ $L* & . L (7'='=j,X\XfXfht'u $J (? ;; + +w 6"9$++:Z:Z"[ $7
%
$}}C$2H2HLL26//4..t
%
%
! \ *k));;;;FFH kk+. L((rb )NNN)rX rY rZ r# intrM r torchTensortupler r
LongTensorr r r r` ra s @rR rj rj
s M} M M0 %0A6R
26+/59
1) ||1) #5<< #= >1) ! . 1)
"% 1) !!1!12
1) +,1)
u||Xell 3XeELL>Q5RR S1) S1)rb rj c e Zd Zy)
Exaone4MLPNre r9 rb rR r r X rf rb r c e Zd Zy)Exaone4DecoderLayerNre r9 rb rR r r \ rf rb r c e Zd ZeZdgZy)Exaone4PreTrainedModelr N)rX rY rZ r# config_class_no_split_modulesr9 rb rR r r ` s L./rb r c e Zd Zdef fd Ze d
dej deej deej dee
deej dee d eej d
e
e d eeef fd Z xZS ) Exaone4Modelrk c $ t | | t j t |j
D cg c] }t
|| c} | _ t |j |j | _ | j y c c}w )Nrp )
rL rM r
ModuleListrJ r< r r. rd r; rC r/ post_initr~ s rR rM zExaone4Model.__init__f so
mmEJ6KcKcEd e
3 e
#6#5#56;N;NO
fs B
r) r, position_idsr& r* rD r rO r c |d u |d uz r t d || j | }|r|t | j }|F||j nd} t j | | |j d z |j }||j d }t |x}
t sF| j |||||d} dt d i | i}
d| j j v rt d i | |
d< |} | j | | }
t! | j" D ]1 \ }}| j j | } || f|
|
| ||||d |} 3 | j% | } t' | |r|
S d
S ) Nz:You must specify exactly one of input_ids or inputs_embeds)rk r r1 )device)rk input_embedsr, r r& r r3 r2 )r r, r r& rD r )last_hidden_stater& r9 )
ValueErrorr- r
rk get_seq_lengthr aranger r unsqueeze
isinstancedictr
rI r
rotary_emb enumerater. r/ r )rN r) r, r r& r* rD r rO past_seen_tokenscausal_mask_mapping mask_kwargsr+ r rP
decoder_layer
layer_types rR r zExaone4Model.forwardp s
-t"; < YZ Z - -i 8M 0 *$++ >O !CRC^ = = ?de "\\ "2]5H5H5K"KTaThTh N )33A6L ?-F ++ -"0"0#2 ,
K !"4"C{"C# #dkk&=&==;\;k_j;k#$78%
"oom\ J )$++ 6 A} 003J ) $7 2: > ) /# - M -0
& +/8O
>B
rb )NNNNNNN)rX rY rZ r# rM r r r r r r FloatTensorboolr r r r r r r` ra s @rR r r e s } '+1537+/59$(59E
##E
! .E
u / /0 E
"% E
1 12
E
D>E
!!1!12E
+,E
u-- .E
E
rb r c , e Zd Z d
deej
deej deej
dee deej deej
dee deej
d e
e ej f d
e e
d ef fd
Z xZS )Exaone4ForCausalLMr) r, r r& r* labelsrD r logits_to_keeprO r c
8 t | d||||||||| d |
y)u
labels (`torch.LongTensor` of shape `(batch_size, sequence_length)`, *optional*):
Labels for computing the masked language modeling loss. Indices should either be in `[0, ...,
config.vocab_size]` or -100 (see `input_ids` docstring). Tokens with indices set to `-100` are ignored
(masked), the loss is only computed for the tokens with labels in `[0, ..., config.vocab_size]`.
Example:
```python
>>> from transformers import AutoModelForCausalLM, AutoTokenizer
>>> model = AutoModelForCausalLM.from_pretrained("LGAI-EXAONE/EXAONE-4.0-Instruct")
>>> tokenizer = AutoTokenizer.from_pretrained("LGAI-EXAONE/EXAONE-4.0-Instruct")
>>> prompt = "Explain how wonderful you are"
>>> messages = [
{"role": "system", "content": "You are a helpful assistant."},
{"role": "user", "content": prompt}
]
>>> input_ids = tokenizer.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=True,
return_tensors="pt",
enable_thinking=False,
)
>>> output = model.generate(input_ids, max_new_tokens=128)
>>> tokenizer.decode(output[0], skip_special_tokens=False)
"[|system|]\nYou are a helpful assistant.[|endofturn|]\n[|user|]\nExplain how wonderful you are[|endofturn|]\n[|assistant|]\n\n\n\n\nOh, thank you for such a kind and lovely question! 😊 \n\nI’m *so* wonderful because I’m here to make your life easier, brighter, and more fun! Whether you need help with: \n\n✨ **Learning** – I can explain anything, from quantum physics to baking the perfect cake! \n💡 **Creativity** – Need a poem, story, or a wild idea? I’ve got you covered! \n🤖 **Problem-solving** – Stuck on a math problem or a tricky decision? I’ll help you figure it out"
```
NOTE: `EXAONE-4.0-Instruct` is a placeholder model ID. The exact model ID will be updated in the future.) r) r, r r& r* r rD r r Nr9 )rL r ) rN r) r, r r& r* r rD r r rO rQ s rR r zExaone4ForCausalLM.forward s< Z
)% +' ))
rb ) NNNNNNNNr )rX rY rZ r r r r r r r r r r r r r r` ra s @rR r r s 151537+/59-1$(59348
E , ,-8
! .8
u / /0 8
"% 8
1 12
8
))*8
D>8
!!1!128
c5<< /08
+,8
8
8
rb r c e Zd Zy) Exaone4ForSequenceClassificationNre r9 rb rR r r rf rb r c e Zd Zy) Exaone4ForTokenClassificationNre r9 rb rR r r rf rb r c e Zd Zy)Exaone4ForQuestionAnsweringNre r9 rb rR r r rf rb r )r# r r r r r r )?r[ typingr r r r r transformers.utils.genericr cache_utilsr r
configuration_utilsr r
masking_utilsr
r modeling_outputsr r modeling_utilsr processing_utilsr utilsr r utils.deprecationr llama.modeling_llamar r r r r r r r r r olmo2.modeling_olmo2r! r"
get_loggerrX logger_CHECKPOINT_FOR_DOC_CONFIG_FOR_DOCr# rd rh Modulerj r r r r r r r r __all__r9 rb rR r s , , 9 . J R 6 & 1 ?
H %7 !C
$ C
L \ 1 K)ryy K)\ + 0 1 0
Q
): Q
h9
) 9
x 'E $? "; rb