gated_delta_net_pruning_mixin

Classes

GatedDeltaNetLayerDescriptor

GatedDeltaNetPruningMixIn

class GatedDeltaNetLayerDescriptor

Bases: LayerDescriptor

__init__(target_name='linear_attn', gdn_prefix_name='model.language_model.layers.{layer_idx}.linear_attn')
Parameters:
  • target_name (str)

  • gdn_prefix_name (str)

Return type:

None

canonical_score_key(method, block_idx, fallback)
Parameters:
  • method (str)

  • block_idx (int | None)

  • fallback (str)

Return type:

str

gdn_prefix(layer_idx)
Parameters:

layer_idx (int)

Return type:

str

gdn_prefix_name: str = 'model.language_model.layers.{layer_idx}.linear_attn'
module_name_regex()
Return type:

str

target_name: str = 'linear_attn'
class GatedDeltaNetPruningMixIn

Bases: PruningMixIn

__init__(layer_descriptor)
Parameters:

layer_descriptor (GatedDeltaNetLayerDescriptor)

supported_hooks()
Return type:

list[type[ForwardHook]]