Functions — github.com/facebookresearch/xformers

CodeHub MCPcopy Create free account

hub / github.com/facebookresearch/xformers / functions

Functions1,445 in github.com/facebookresearch/xformers

⨍ Functions 1,445◇ Types & classes 346↳ Endpoints 17

FunctionSTABLE_TORCH_LIBRARY_FRAGMENT

xformers/csrc/sparse24/sparse24.cpp:3

xformers/csrc/sparse24/static_sort.h:24

FunctionTORCH_LIBRARY_FRAGMENT

xformers/csrc/attention/hip_fmha/ck_fmha_test.cpp:21

FunctionTORCH_LIBRARY_IMPL

xformers/csrc/attention/hip_decoder/attention_forward_splitk.cpp:312

FunctionTORCH_LIBRARY_IMPL

xformers/csrc/attention/hip_fmha/attention_forward_generic_ck_tiled.cpp:523

FunctionTORCH_LIBRARY_IMPL

xformers/csrc/attention/hip_fmha/attention_backward_generic_ck_tiled.cpp:630

FunctionTORCH_LIBRARY_IMPL

xformers/csrc/attention/hip_fmha/attention_ck_rand_uniform.cpp:93

MethodTile4x4Packed

xformers/csrc/sparse24/sparse24_pack.h:186

MethodTile8x8Masks

xformers/csrc/sparse24/sparse24_pack.h:30

MethodTile8x8Meta

xformers/csrc/sparse24/sparse24_pack.h:266

MethodTileValueOrdered1d

xformers/csrc/sparse24/warp_tensor.h:324

MethodTileValueOrderedT

xformers/csrc/sparse24/compute_sparse_tile.h:47

MethodWarpTensor<Element, kRows, kCols> sparsify_dense

xformers/csrc/sparse24/warp_tensor.h:328

MethodWarpTensor<ElementOut, kRows, kCols> to

xformers/csrc/sparse24/warp_tensor.h:130

(self, ctx, func, *args, **kwargs)

xformers/checkpoint.py:466

xformers/fwbw_overlap.py:303

xformers/fwbw_overlap.py:331

(self, *args: Optional[torch.Tensor])

xformers/ops/swiglu_op.py:86

(self, *args, **kwargs)

xformers/ops/swiglu_op.py:97

(self, *args, **kwargs)

xformers/ops/swiglu_op.py:102

Method__enter__

xformers/profiler/profiler.py:42

Method__enter__

xformers/profiler/profiler.py:137

Method__enter__

xformers/profiler/profiler.py:173

Method__enter__

xformers/profiler/profiler_dcgm_impl.py:152

Method__enter__

xformers/profiler/profiler_dcgm.py:32

Method__enter__

tests/multiprocessing_utils.py:61

Method__enter__

tests/multiprocessing_utils.py:146

(self, exc_type, exc_val, exc_tb)

xformers/profiler/profiler.py:45

(self, exc_type, exc_val, exc_tb)

xformers/profiler/profiler.py:141

(self, exc_type, exc_val, exc_tb)

xformers/profiler/profiler.py:188

(self, exc_type, exc_val, exc_tb)

xformers/profiler/profiler_dcgm_impl.py:176

(self, exc_type, exc_val, exc_tb)

xformers/profiler/profiler_dcgm.py:38

(self, exc_type, exc_val, exc_tb)

tests/multiprocessing_utils.py:64

(self, exc_type, exc_val, exc_tb)

tests/multiprocessing_utils.py:158

Method__getattr__

(self, name: str)

tests/multiprocessing_utils.py:58

Method__getitem__

(self, step: Step)

xformers/benchmarks/benchmark_sequence_parallel_fused.py:46

(self, *args, **kwargs)

xformers/checkpoint.py:99

(self, policy_fn, storage, allow_cache_entry_mutation)

xformers/checkpoint.py:121

(self, num_runs: int = 10)

xformers/checkpoint.py:210

(self, optim_output: torch.Tensor)

xformers/checkpoint.py:462

(self, mod, memory_budget=None, policy_fn=None)

xformers/checkpoint.py:476

xformers/fwbw_overlap.py:25

(self, event: Union[EventHandle, None] = None)

xformers/fwbw_overlap.py:32

( self, event_overlap: Union[EventOverlap, None], device: torch.device, name:

xformers/fwbw_overlap.py:83

(self, build_info: Optional[_BuildInfo])

xformers/_cpp_lib.py:66

( self, device: torch.device, group: dist.ProcessGroup, )

xformers/ops/sequence_parallel_fused_ops.py:68

( self, in_features: int, out_features: List[int], *, process_group: t

xformers/ops/modpar_layers.py:46

(self, dim: int, include_weight: bool = True, eps: float = 1e-6)

xformers/ops/rmsnorm.py:91

(self, op, packed_weights: bool, name: str, constraints)

xformers/ops/swiglu_op.py:75

Create a SwiGLU module Args: in_features (int): Number of features of the input hidden_features (int): Number of hidd

xformers/ops/swiglu_op.py:308

Internal use only

xformers/ops/fmha/merge_training.py:126

not used, only here for backward compatibility

xformers/ops/fmha/attn_bias.py:153

(self, bias: torch.Tensor)

xformers/ops/fmha/attn_bias.py:363

xformers/triton/vararg_kernel.py:43

(self, *args, **kwargs)

xformers/triton/vararg_kernel.py:120

(self, values, layout)

xformers/sparse/blocksparse_tensor.py:91

(self, e: torch._C._autograd._KinetoEvent)

xformers/profiler/profile_analyzer.py:15

(self, main_profiler: "_Profiler")

xformers/profiler/profiler.py:38

(self, main_profiler: "_Profiler")

xformers/profiler/profiler.py:62

(self, main_profiler: "_Profiler")

xformers/profiler/profiler.py:165

( self, output_dir: str, schedule: Sequence[Tuple[Any, int, int]], module: Opt

xformers/profiler/profiler.py:226

Args: main_profiler: The main profiler object. gpus_to_profile: A tuple of integers representing the GPUs to profile.

xformers/profiler/profiler_dcgm_impl.py:21

( self, main_profiler: "_Profiler", gpus_to_profile=None, fiel

xformers/profiler/profiler_dcgm.py:23

( self, B: int, Mq: int, Mkv: int, Hq: int, Hkv: int,

xformers/benchmarks/benchmark_attn_decoding.py:78

( self, B: int, Mq: int, Mkv: int, Hq: int, Hkv: int,

xformers/benchmarks/benchmark_attn_decoding.py:171

( self, B: int, Mq: int, Mkv: int, Hq: int, Hkv: int,

xformers/benchmarks/benchmark_attn_decoding.py:250

(self, dtype, scaling: bool, shape, bw: bool)

xformers/benchmarks/benchmark_indexing.py:51

(self, dtype, batches, D, keep_ratio, bw: bool)

xformers/benchmarks/benchmark_indexing.py:107

( self, B_in_hidden_out_ft: Tuple[int, int, int, int], dtype, bias: bool, bw: bool )

xformers/benchmarks/benchmark_sp24.py:116

(self, blocks, policy_fn)

tests/test_checkpoint.py:279

(self, dim: int, include_weight: bool = True, eps: float = 1e-6)

tests/test_rmsnorm.py:27

( self, in_features: int, hidden_features=None, out_features=None, act

tests/test_sparsity24.py:959

tests/multiprocessing_utils.py:18

(self, world_size: int)

tests/multiprocessing_utils.py:143

Initializes the Tokenizer with a SentencePiece model. Args: model_path (str): The path to the SentencePiece model file.

examples/llama_inference/tokenizer.py:14

( self, args: GenArgs, model_args: fast.ModelArgs, model: fast.Transformer,

examples/llama_inference/generate.py:87

( self, dim: int, hidden_dim: int, multiple_of: int, ffn_dim_multiplie

examples/llama_inference/model.py:155

(self, args: ModelArgs, layer_index: int)

examples/llama_inference/model.py:210

(self, args: ModelArgs)

examples/llama_inference/model.py:267

examples/llama_inference/stats.py:32

( cls: "type[EventOverlapHolder]", event_overlap: Union[EventOverlap, None], device: t

xformers/fwbw_overlap.py:68

( cls, shape, packed: torch.Tensor, meta: torch.Tensor, packed_t: torc

xformers/ops/sp24.py:300

Creates a new ScaledTensor subclass instance. Parameters: - data: The underlying quantized tensor (e.g., int8, int4).

xformers/ops/fmha/common.py:64

Method__post_init__

xformers/fwbw_overlap.py:293

Method__post_init__

xformers/ops/fmha/attn_bias.py:224

Method__post_init__

xformers/ops/fmha/attn_bias.py:339

Method__post_init__

xformers/ops/fmha/attn_bias.py:1049

Method__post_init__

xformers/ops/fmha/attn_bias.py:1876

Method__post_init__

xformers/ops/fmha/attn_bias.py:1942

xformers/fwbw_overlap.py:97

xformers/ops/sp24.py:326

Custom string representation for ScaledTensor.

xformers/ops/fmha/common.py:117

xformers/sparse/blocksparse_tensor.py:104

xformers/fwbw_overlap.py:297

xformers/_cpp_lib.py:56

xformers/_cpp_lib.py:69

xformers/ops/swiglu_op.py:89

xformers/benchmarks/benchmark_sequence_parallel_fused.py:37

Method__tensor_flatten__

xformers/fwbw_overlap.py:94

← previous next →601–700 of 1,445, ranked by callers