easydel.modules.exaone.modeling_exaone

class easydel.modules.exaone.modeling_exaone.ExaoneAttention(*args: Any, **kwargs: Any)

Bases: Module

Wrapper around ExaoneAttentionInner to wire it into decoder layers.

class easydel.modules.exaone.modeling_exaone.ExaoneAttentionInner(*args: Any, **kwargs: Any)

Bases: UnifiedAttention

Exaone attention with partial RoPE.

projection_mapping: ClassVar[dict[str, str]] = {'key_projection': 'k_proj', 'output_projection': 'out_proj', 'qkv_projection': 'qkv_proj', 'query_projection': 'q_proj', 'value_projection': 'v_proj'}

class easydel.modules.exaone.modeling_exaone.ExaoneDecoderLayer(*args: Any, **kwargs: Any)

Bases: Module

Single Exaone decoder block combining attention and gated MLP.

class easydel.modules.exaone.modeling_exaone.ExaoneForCausalLM(*args: Any, **kwargs: Any)

Bases: BaseCausalLMModule[ExaoneModel, ExaoneConfig]

Exaone model with a language modeling head for causal language modeling tasks.

class easydel.modules.exaone.modeling_exaone.ExaoneForSequenceClassification(*args: Any, **kwargs: Any)

Bases: BaseSequenceClassificationModule[ExaoneModel, ExaoneConfig]

Exaone model with a Sequence Classification head.

class easydel.modules.exaone.modeling_exaone.ExaoneGatedMLP(*args: Any, **kwargs: Any)

Bases: Module

Gated feed-forward block used inside Exaone decoder layers.

class easydel.modules.exaone.modeling_exaone.ExaoneModel(*args: Any, **kwargs: Any)

Bases: EasyDeLBaseModule

Decoder-only Exaone transformer composed of embedding, stacked blocks, and final norm.

property frequencies

Retrieves or computes the frequency components (e.g., for RoPE) from the configuration.

Uses self.config.get_basic_frequencies() and caches the result.

Returns

The frequency components, potentially cached.

Return type

jnp.ndarray

get_decoder()

Returns the decoder part of the model’s graph definition.

get_embedding()

Returns the embedding layer of the module.

get_encoder()

Returns the encoder part of the model’s graph definition. Decoder-Only models don’t have an encoder.

get_lm_head()

Returns the language model head of the module. Base Models don’t have a Language Model Head.