Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/fla-org/flash-linear-attention
/ functions
Functions
2,989 in github.com/fla-org/flash-linear-attention
⨍
Functions
2,989
◇
Types & classes
403
↳
Endpoints
219
Method
__init__
(self, *inputs, **kwargs)
fla/models/rwkv6/modeling_rwkv6.py:203
Method
__init__
(self, config: RWKV6Config)
fla/models/rwkv6/modeling_rwkv6.py:249
Method
__init__
(self, config)
fla/models/rwkv6/modeling_rwkv6.py:345
Method
__init__
( self, attn_mode: str = "chunk", hidden_size: int = 2048, expand_k: float = 0
fla/models/rwkv6/configuration_rwkv6.py:18
Method
__init__
(self, *inputs, **kwargs)
fla/models/gla/modeling_gla.py:185
Method
__init__
(self, config: GLAConfig)
fla/models/gla/modeling_gla.py:237
Method
__init__
(self, config)
fla/models/gla/modeling_gla.py:348
Method
__init__
( self, hidden_size: int = 2048, expand_k: float = 0.5, expand_v: float = 1.,
fla/models/gla/configuration_gla.py:18
Method
__init__
(self, config)
fla/models/log_linear_mamba2/modeling_log_linear_mamba2.py:264
Method
__init__
(self, config)
fla/models/log_linear_mamba2/modeling_log_linear_mamba2.py:412
Method
__init__
( self, residual_in_fp32: bool = False, chunk_size: int = 64, attnres_block_si
fla/models/log_linear_mamba2/configuration_log_linear_mamba2.py:15
Method
__init__
( self, head_dim: int = 64, vocab_size: int = 32000, hidden_size: int = 2048,
fla/models/mamba3/configuration_mamba3.py:25
Method
__init__
(self, config: Mamba3Config)
fla/models/mamba3/modeling_mamba3.py:162
Method
__init__
(self, config: Mamba3Config)
fla/models/mamba3/modeling_mamba3.py:252
Method
__init__
( self, hidden_size: int = 2048, num_hidden_layers: int = 24, attn_mode: str =
fla/models/lightnet/configuration_lightnet.py:18
Method
__init__
(self, *inputs, **kwargs)
fla/models/lightnet/modeling_lightnet.py:177
Method
__init__
(self, config: LightNetConfig)
fla/models/lightnet/modeling_lightnet.py:229
Method
__init__
(self, config)
fla/models/lightnet/modeling_lightnet.py:341
Method
__init__
(self, *inputs, **kwargs)
fla/models/transformer/modeling_transformer.py:180
Method
__init__
(self, config: TransformerConfig)
fla/models/transformer/modeling_transformer.py:229
Method
__init__
(self, config)
fla/models/transformer/modeling_transformer.py:350
Method
__init__
( self, hidden_size: int = 2048, num_hidden_layers: int = 24, num_heads: int =
fla/models/transformer/configuration_transformer.py:18
Method
__init__
( self, attn_mode: str = "chunk", hidden_size: int = 2048, conv_size: int = 4,
fla/models/comba/configuration_comba.py:17
Method
__init__
(self, *inputs, **kwargs)
fla/models/comba/modeling_comba.py:180
Method
__init__
(self, config: CombaConfig)
fla/models/comba/modeling_comba.py:246
Method
__init__
(self, config)
fla/models/comba/modeling_comba.py:357
Method
__init__
(self, *inputs, **kwargs)
fla/models/nsa/modeling_nsa.py:180
Method
__init__
(self, config: NSAConfig)
fla/models/nsa/modeling_nsa.py:234
Method
__init__
(self, config)
fla/models/nsa/modeling_nsa.py:350
Method
__init__
( self, hidden_size: int = 2048, num_hidden_layers: int = 24, num_heads: int =
fla/models/nsa/configuration_nsa.py:18
Method
__init__
( self, attn_mode: str = "chunk", hidden_size: int = 2048, expand_v: float = 1
fla/models/kda/configuration_kda.py:15
Method
__init__
(self, *inputs, **kwargs)
fla/models/kda/modeling_kda.py:193
Method
__init__
(self, config: KDAConfig)
fla/models/kda/modeling_kda.py:260
Method
__init__
(self, config)
fla/models/kda/modeling_kda.py:375
Method
__init__
(self, *inputs, **kwargs)
fla/models/gsa/modeling_gsa.py:186
Method
__init__
(self, config: GSAConfig)
fla/models/gsa/modeling_gsa.py:238
Method
__init__
(self, config)
fla/models/gsa/modeling_gsa.py:349
Method
__init__
( self, hidden_size: int = 2048, gate_logit_normalizer: int | None = 8, clamp_
fla/models/gsa/configuration_gsa.py:18
Method
__iter__
(self)
legacy/training/flame/data.py:60
Method
__iter__
(self)
fla/models/utils.py:214
Method
__iter__
(self)
fla/models/utils.py:416
Method
__len__
(self)
fla/models/utils.py:217
Method
__len__
(self)
fla/models/utils.py:420
Method
__repr__
(self)
fla/modules/layernorm.py:161
Method
__repr__
(self)
fla/modules/layernorm.py:1003
Method
__repr__
(self)
fla/modules/layernorm.py:1063
Method
__repr__
(self)
fla/modules/layernorm.py:1120
Method
__repr__
(self)
fla/modules/layernorm.py:1278
Method
__repr__
(self)
fla/modules/layernorm.py:1341
Method
__repr__
(self)
fla/modules/layernorm.py:1400
Method
__repr__
(self)
fla/modules/fused_bitlinear.py:583
Method
__repr__
(self)
fla/modules/fused_norm_gate.py:962
Method
__repr__
(self)
fla/modules/fused_norm_gate.py:1025
Method
__repr__
(self)
fla/modules/fused_norm_gate.py:1121
Method
__repr__
(self)
fla/modules/fused_norm_gate.py:1198
Method
__repr__
(self)
fla/modules/rotary.py:390
Method
__repr__
(self)
fla/modules/feature_map.py:94
Method
__repr__
(self)
fla/layers/rwkv6.py:251
Method
__repr__
(self)
fla/layers/rwkv6.py:319
Method
__repr__
(self)
fla/layers/rwkv6.py:353
Function
_add_kernel
( a_ptr, b_ptr, out_ptr, n_elements, BLOCK: tl.constexpr, )
fla/modules/backends/triton_ascend/causal_conv1d.py:192
Method
_apply
(self, module: nn.Module, device_mesh: DeviceMesh)
fla/modules/layernorm.py:1465
Method
_apply
(self, module: nn.Module, device_mesh: DeviceMesh)
fla/modules/parallel.py:37
Method
_apply
(self, module: nn.Module, device_mesh: DeviceMesh)
fla/modules/mlp.py:134
Method
_apply
(self, module: nn.Module, device_mesh: DeviceMesh)
fla/modules/fused_linear_cross_entropy.py:760
Method
_compute_inv_freq
(self, device=None)
fla/layers/yoco.py:54
Function
_config_deltaformer
()
fla/ops/deltaformer/parallel.py:125
Function
_default_alloc_fn
(size: int, alignment: int, stream: int | None)
fla/utils/_device.py:143
Function
_fwdbwd_fn
(inputs=inputs, do=do)
benchmarks/ops/run.py:292
Function
_guarded_empty
Create a tensor filled with NaN instead of uninitialized values.
tests/conftest.py:62
Function
_guarded_empty_like
Create a tensor filled with NaN instead of uninitialized values.
tests/conftest.py:82
Function
_guarded_new_empty
Create a tensor filled with NaN instead of uninitialized values.
tests/conftest.py:104
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/mla/modeling_mla.py:172
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/abc/modeling_abc.py:186
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = False, num_residua
fla/models/wall_transformer/modeling_wall_transformer.py:186
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/moba/modeling_moba.py:172
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/linear_attn/modeling_linear_attn.py:186
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = False, num_residua
fla/models/mom/modeling_mom.py:277
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/delta_net/modeling_delta_net.py:186
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/yoco/modeling_yoco.py:333
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/hgrn2/modeling_hgrn2.py:183
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/mesa_net/modeling_mesa_net.py:184
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/gated_deltaproduct/modeling_gated_deltaproduct.py:185
Method
_init_weights
Initialize the weights.
fla/models/mamba2/modeling_mamba2.py:146
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, )
fla/models/rodimus/modeling_rodimus.py:260
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = True, num_residual
fla/models/rwkv7/modeling_rwkv7.py:234
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = False, num_residua
fla/models/bitnet/modeling_bitnet.py:223
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/deltaformer/modeling_deltaformer.py:164
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/hgrn/modeling_hgrn.py:182
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/gated_deltanet/modeling_gated_deltanet.py:184
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/retnet/modeling_retnet.py:186
Method
_init_weights
Initialize the weights.
fla/models/mamba/modeling_mamba.py:110
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = False, num_residua
fla/models/path_attn/modeling_path_attention.py:181
Method
_init_weights
Initialize the weights.
fla/models/samba/modeling_samba.py:187
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/raven/modeling_raven.py:198
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = False, num_residua
fla/models/forgetting_transformer/modeling_forgetting_transformer.py:182
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = False, num_residua
fla/models/parallax/modeling_parallax.py:170
Method
_init_weights
( self, module: nn.Module, rescale_prenorm_residual: bool = True, num_residual
fla/models/rwkv6/modeling_rwkv6.py:206
Method
_init_weights
( self, module: nn.Module, prenorm_residual_strategy: str | None = None, num_r
fla/models/gla/modeling_gla.py:188
Method
_init_weights
Initialize the weights.
fla/models/log_linear_mamba2/modeling_log_linear_mamba2.py:162
← previous
next →
1,201–1,300 of 2,989, ranked by callers