easydel.modules.deepseek_v2.modeling_deepseek_flax

class easydel.modules.deepseek_v2.modeling_deepseek_flax.DeepseekV2Attention(*args: Any, **kwargs: Any)

Bases: AttentionModule

class easydel.modules.deepseek_v2.modeling_deepseek_flax.DeepseekV2DecoderLayer(*args: Any, **kwargs: Any)

Bases: Module

class easydel.modules.deepseek_v2.modeling_deepseek_flax.DeepseekV2ForCausalLM(*args: Any, **kwargs: Any)

Bases: EasyDeLBaseModule

DeepseekV2 model with a language modeling head for causal language modeling tasks.

This model extends the base DeepseekV2Model by adding a linear language modeling head on top of the transformer model. It’s designed for generative tasks and can be used for text generation.

class easydel.modules.deepseek_v2.modeling_deepseek_flax.DeepseekV2MLP(*args: Any, **kwargs: Any)

Bases: Module

class easydel.modules.deepseek_v2.modeling_deepseek_flax.DeepseekV2MoE(*args: Any, **kwargs: Any)

Bases: Module

class easydel.modules.deepseek_v2.modeling_deepseek_flax.DeepseekV2Model(*args: Any, **kwargs: Any)

Bases: EasyDeLBaseModule

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

class easydel.modules.deepseek_v2.modeling_deepseek_flax.MoEGate(*args: Any, **kwargs: Any)

Bases: Module

easydel.modules.deepseek_v2.modeling_deepseek_flax.apply_rotary_pos_emb(q, k, cos, sin, position_ids, unsqueeze_dim=1)

easydel.modules.deepseek_v2.modeling_deepseek_flax.init_deepseek_rotary_embedding(dim, max_position_embeddings=2048, base=10000, method: Literal['linear', 'yarn', 'dynamic', None] = None, kwargs: Optional[dict] = None)

easydel.modules.deepseek_v2.modeling_deepseek_flax.rotate_half(x)

easydel.modules.deepseek_v2.modeling_deepseek_flax.yarn_find_correction_dim(num_rotations, dim, base=10000, max_position_embeddings=2048)

easydel.modules.deepseek_v2.modeling_deepseek_flax.yarn_find_correction_range(low_rot, high_rot, dim, base=10000, max_position_embeddings=2048)

easydel.modules.deepseek_v2.modeling_deepseek_flax.yarn_get_mscale(scale=1.0, mscale=1.0)

easydel.modules.deepseek_v2.modeling_deepseek_flax.yarn_linear_ramp_mask(min, max, dim)