Functionsigmoid_bwd_kernel(
x, dy, dx,
stride_x_row,
stride_dy_row,
stride_dx_row,
T,
D: tl.constexpr,
B: tl
fla/modules/activations.py:126
Functionsigmoid_bwd_kernel(
x, dy, dx,
T,
D: tl.constexpr,
stride_x_row,
stride_dy_row,
stride_dx_row,
B: tl
fla/modules/backends/triton_ascend/activations.py:103
Functionsigmoid_fwd_kernel(
x, y,
stride_x_row,
stride_y_row,
T,
D: tl.constexpr,
B: tl.constexpr,
)
fla/modules/activations.py:98
Functionsigmoid_fwd_kernel(
x, y,
T,
D: tl.constexpr,
stride_x_row,
stride_y_row,
B: tl.constexpr,
)
fla/modules/backends/triton_ascend/activations.py:73
Functionsigmoidglu_fwd_kernel(
x, y, z,
stride_x_row,
stride_y_row,
stride_z_row,
T,
D: tl.constexpr,
B: tl.con
fla/modules/activations.py:772
Functionsigmoidglu_fwdbwd_kernel(
x, y, g, dx, dy, z,
stride_x_row,
stride_y_row,
stride_g_row,
stride_dx_row,
stride_
fla/modules/activations.py:805
Functionsolve_tril_16x16_kernel(
A,
Ai,
cu_seqlens,
chunk_indices,
T,
H: tl.constexpr,
BT: tl.constexpr,
USE_
fla/ops/utils/solve_tril.py:37
Functionswiglu_fwd_kernel(
x, y, z,
stride_x_row,
stride_y_row,
stride_z_row,
T,
D: tl.constexpr,
B: tl.con
fla/modules/activations.py:571
Functionswiglu_fwd_kernel(
x, y, z,
T,
D: tl.constexpr,
stride_x_row,
stride_y_row,
stride_z_row,
B: tl.con
fla/modules/backends/triton_ascend/activations.py:275
Methodswiglu_fwdbwd(self, x, y, g, use_weight=False, output_contiguous=False)
fla/modules/backends/triton_ascend/__init__.py:167
Functionswiglu_fwdbwd_kernel(
x, y, g, dx, dy, z,
stride_x_row,
stride_y_row,
stride_g_row,
stride_dx_row,
stride_
fla/modules/activations.py:604
Functionswiglu_fwdbwd_kernel(
x, y, g, dx, dy, z,
T,
D: tl.constexpr,
stride_x_row,
stride_y_row,
stride_g_row,
fla/modules/backends/triton_ascend/activations.py:312
Functionswish_bwd_kernel(
x, dy, dx,
stride_x_row,
stride_dy_row,
stride_dx_row,
T,
D: tl.constexpr,
B: tl
fla/modules/activations.py:369
Functionswish_bwd_kernel(
x, dy, dx,
T,
D: tl.constexpr,
stride_x_row,
stride_dy_row,
stride_dx_row,
B: tl
fla/modules/backends/triton_ascend/activations.py:241
Functionswish_fwd_kernel(
x, y,
stride_x_row,
stride_y_row,
T,
D: tl.constexpr,
B: tl.constexpr,
)
fla/modules/activations.py:341
Functionswish_fwd_kernel(
x, y,
T,
D: tl.constexpr,
stride_x_row,
stride_y_row,
B: tl.constexpr,
)
fla/modules/backends/triton_ascend/activations.py:210
Functiontest_attnres(
L: int,
B: int,
T: int,
D: int,
scale: float,
fuse_output_norm: bool,
dtype: tor
tests/ops/test_attnres.py:42
Functiontest_channel_mixing_gradients(B, T, n_embd, dim_ffn, dtype, inplace, xprevdim)
tests/ops/test_rwkv7.py:34
Functiontest_chunk(
B: int,
T: int,
H: int,
D: int,
scale: float,
use_qk_l2norm_in_kernel: bool,
dty
tests/ops/test_delta.py:34
Functiontest_chunk(
B: int,
T: int,
H: int,
D: int,
scale: float,
gate_logit_normalizer: float,
mask
tests/ops/test_comba.py:138
Functiontest_chunk(
B: int,
T: int,
H: int,
D: int,
scale: float,
num_householder: int,
use_qk_l2nor
tests/ops/test_delta_product.py:37
Functiontest_chunk(
B: int,
T: int,
H: int,
D: int,
mask_p: float,
gate_logit_normalizer: float,
saf
tests/ops/test_dplr_delta.py:288
Functiontest_chunk(
B: int,
T: int,
H: int,
HV: int,
D: int,
scale: float,
gate_logit_normalizer: fl
tests/ops/test_kda.py:532
Functiontest_chunk(
B: int,
T: int,
H: int,
D: int,
M: int,
gate_logit_normalizer: float,
dtype: tor
tests/ops/test_gsa.py:218