MCPcopy Create free account

hub / github.com/KarelPeeters/Kyanite / functions

Functions1,306 in github.com/KarelPeeters/Kyanite

⨍ Functions 1,306◇ Types & classes 523
↓ 9 callersMethodsynchronize
(&self)
kn-cuda-sys/src/wrapper/event.rs:27
↓ 9 callersMethodvalues
Iterate over the values in this graph, in topological order, which means that nodes will only be visited after all of their inputs have been visited.
kn-graph/src/graph.rs:391
↓ 8 callersFunction__hneg
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:3826
↓ 8 callersMethodas_single_const
Returns `Some(f)` if `value` is effectively a constant with every element equal to `f`.
kn-graph/src/graph.rs:490
↓ 8 callersMethodeval
(&self, batch_size: usize)
kn-graph/src/shape.rs:94
↓ 8 callersFunctionint
* Conversion operator to \p int data type. * Using round-toward-zero rounding mode. * * See __half2int_rz(__half) for further details
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:491
↓ 8 callersMethodproduct
(iter: I)
kn-graph/src/shape.rs:504
↓ 8 callersMethodreduce
(&mut self, input: Value, axes: Vec<usize>, op: ReduceOp)
kn-graph/src/graph.rs:1156
↓ 8 callersFunctionrng_vec
(len: usize, rng: &mut impl Rng)
kn-cuda-eval/tests/root/tensor_utils.rs:23
↓ 8 callersMethodtake_float
(&mut self, key: &str)
kn-graph/src/onnx/inputs.rs:178
↓ 7 callersMethodcopy_linear_to_host
(&self, buffer: &mut [u8])
kn-cuda-sys/src/wrapper/mem/device.rs:128
↓ 7 callersMethoddevice
(&self)
kn-cuda-eval/src/planner.rs:1006
↓ 7 callersMethoddtype
(self)
kn-graph/src/dtype.rs:244
↓ 7 callersMethodevaluate
TODO accept views as inputs? introduce DView struct/alias?
kn-cuda-eval/src/executor.rs:117
↓ 7 callersMethodfinish
(self)
kn-cuda-sys/src/wrapper/rtc/args.rs:47
↓ 7 callersMethodflip
(&self, axis: usize)
kn-cuda-eval/src/shape.rs:96
↓ 7 callersFunctionmake_cuDoubleComplex
kn-cuda-sys/doc_headers/cuda_include/cuComplex.h:198
↓ 7 callersFunctionmake_cuFloatComplex
kn-cuda-sys/doc_headers/cuda_include/cuComplex.h:89
↓ 7 callersMethodmaybe_take
(&mut self, key: &str, ty: AttributeType)
kn-graph/src/onnx/inputs.rs:110
↓ 7 callersMethodmaybe_take_string
(&mut self, key: &str)
kn-graph/src/onnx/inputs.rs:136
↓ 7 callersMethodoffset
(&self, offset_elem: isize, shape: StridedShape)
kn-cuda-eval/src/offset_tensor.rs:53
↓ 7 callersMethodsingle
(mut self)
kn-runtime/examples/stable_diffusion.rs:455
↓ 7 callersMethodsize_bytes
(&self)
kn-cuda-sys/src/wrapper/descriptor.rs:96
↓ 7 callersMethodsplit
(&self, index: usize)
kn-graph/src/shape.rs:206
↓ 7 callersMethodtake_ints
(&mut self, key: &str)
kn-graph/src/onnx/inputs.rs:170
↓ 6 callersFunction__float2half_rd
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1189
↓ 6 callersFunction__float2half_ru
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1206
↓ 6 callersFunction__float2half_rz
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1175
↓ 6 callersMethodadd
(self, rhs: Self)
kn-graph/src/onnx/typed_value.rs:204
↓ 6 callersMethodassert_linear_in_bounds
(&self, len: usize)
kn-cuda-sys/src/wrapper/mem/device.rs:175
↓ 6 callersMethodbit
(&mut self, x: X, y: Y)
kn-cuda-eval/tests/root/elementwise.rs:133
↓ 6 callersMethodconstant
(&mut self, shape: Shape, data: Vec<T>)
kn-graph/src/graph.rs:599
↓ 6 callersMethodis_const_filled_with
Returns whether `value` is effectively a constant with every element equal to `expected`.
kn-graph/src/graph.rs:477
↓ 6 callersMethodlaunch_kernel
( &self, grid_dim: impl Into<Dim3>, block_dim: impl Into<Dim3>, shared_mem_byt
kn-cuda-sys/src/wrapper/rtc/core.rs:251
↓ 6 callersFunctionload_graph_from_onnx_path
Load an [ONNX](https://github.com/onnx/onnx/blob/main/docs/IR.md) file from the given path. If `allow_external` is true, the onnx will be allowed to
kn-graph/src/onnx/mod.rs:28
↓ 6 callersFunctionrange_vec
(len: usize)
kn-cuda-eval/tests/root/tensor_utils.rs:27
↓ 6 callersMethodrank
(&self)
kn-cuda-eval/src/shape.rs:61
↓ 6 callersMethodrecord_event
(&self)
kn-cuda-sys/src/wrapper/handle.rs:149
↓ 6 callersMethodrepeat
Repeat `input` along a given `axis`, `count` times. This starts by emitting the entire tensor before repeating elements, similar to `torch.repeat` or
kn-graph/src/graph.rs:816
↓ 6 callersMethodrepeat_interleave
Repeat elements of `input` along a given `axis`, `count` times. This starts by repeat each element before going to the next one, similar to `torch.re
kn-graph/src/graph.rs:825
↓ 6 callersMethodscalar
(&mut self, value: T)
kn-graph/src/graph.rs:612
↓ 6 callersMethodvisit
(&mut self, old: Value)
kn-graph/src/optimizer/core.rs:36
↓ 5 callersFunction__habs
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:3850
↓ 5 callersFunction__half
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:287
↓ 5 callersFunction__half_raw
* Type cast to \p __half_raw operator. */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:310
↓ 5 callersFunction__internal_bfloat162float
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:1163
↓ 5 callersFunction__internal_float2bfloat16
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:973
↓ 5 callersFunction__internal_float2half
no host functions in NVRTC mode */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:962
↓ 5 callersFunction__internal_half2float
no host functions in NVRTC mode */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1265
↓ 5 callersFunction__nv_bfloat16_raw
* Type cast to \p __nv_bfloat16_raw operator. */
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:276
↓ 5 callersFunctionbroadcast_shape_symmetric
This corresponds to [_multidimensional broadcasting_ in the ONNX spec](https://github.com/onnx/onnx/blob/main/docs/Broadcasting.md#multidirectional-br
kn-graph/src/graph.rs:1431
↓ 5 callersMethodclamp
(&mut self, input: Value, min: T, max: T)
kn-graph/src/graph.rs:1222
↓ 5 callersFunctioncompile_cached_kernel
(key: KernelKey)
kn-cuda-eval/src/autokernel/common.rs:30
↓ 5 callersFunctioncudaDeviceGetAttribute
kn-cuda-sys/doc_headers/cuda_include/cuda_device_runtime_api.h:89
↓ 5 callersMethoddevice
(&self)
kn-cuda-sys/src/wrapper/mem/device.rs:68
↓ 5 callersMethoddevice
(&self)
kn-cuda-eval/src/device_tensor.rs:27
↓ 5 callersMethodeval
(&mut self, inputs: &[DTensor])
kn-runtime/src/lib.rs:120
↓ 5 callersFunctionfill_replacements
(src: &str, replacements: &[(&str, String)])
kn-cuda-eval/src/autokernel/common.rs:48
↓ 5 callersMethodinfo
(self)
kn-graph/src/dtype.rs:97
↓ 5 callersMethodis_hidden_with_uses
(&self, value: Value, users: usize)
kn-graph/src/graph.rs:421
↓ 5 callersMethodmaybe_take_ints
(&mut self, key: &str)
kn-graph/src/onnx/inputs.rs:166
↓ 5 callersMethodoffset_bytes
(self, offset: isize)
kn-cuda-sys/src/wrapper/mem/device.rs:72
↓ 5 callersMethodsoftmax
(&mut self, input: Value, axis: usize)
kn-graph/src/graph.rs:1126
↓ 5 callersMethodto_tensor
(self)
kn-graph/src/dtype.rs:267
↓ 4 callersFunction__internal_float_as_uint
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:947
↓ 4 callersFunction__internal_uint_as_float
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:960
↓ 4 callersMethodas_signed_shape
(&self, graph: &Graph)
kn-graph/src/onnx/typed_value.rs:77
↓ 4 callersMethodas_size
(&self, graph: &Graph)
kn-graph/src/onnx/typed_value.rs:60
↓ 4 callersMethodbatched_mat_mul
(&mut self, left: Value, right: Value)
kn-graph/src/graph.rs:1106
↓ 4 callersMethodcan_fuse
(&self, value: Value)
kn-cuda-eval/src/planner.rs:1010
↓ 4 callersMethodcheck_contains
(&self, value: Value)
kn-graph/src/graph.rs:375
↓ 4 callersFunctioncpu_eval_graph
(graph: &Graph, batch_size: usize, inputs: &[DTensor])
kn-graph/src/cpu.rs:20
↓ 4 callersMethoddummy_zero_inputs
Generate a set of dummy inputs that have the right shapes and dtypes and are all fully zero. This can be useful for some quick testing.
kn-graph/src/graph.rs:1416
↓ 4 callersMethodinto_ptr
(self)
kn-cuda-eval/src/offset_tensor.rs:27
↓ 4 callersMethodlen_bytes
(&self)
kn-cuda-sys/src/wrapper/mem/pinned.rs:62
↓ 4 callersMethodmat_mul
(&mut self, left: Value, right: Value)
kn-graph/src/graph.rs:1063
↓ 4 callersFunctionrepeat
()
kn-cuda-eval/tests/root/graphs.rs:107
↓ 4 callersMethodreplace_all
(&self, axes: &[usize], replacement: Shape)
kn-graph/src/shape.rs:143
↓ 4 callersMethodscalar_dyn
(&mut self, value: DScalar)
kn-graph/src/graph.rs:607
↓ 4 callersMethodset_debug_id
Equivalent to `self[value].debug_id = id`, but that would not work since there is intentionally no implementation of `IndexMut` for `Graph`.
kn-graph/src/graph.rs:578
↓ 4 callersFunctionsimple
()
kn-python/example.py:11
↓ 4 callersMethodtake_int
(&mut self, key: &str)
kn-graph/src/onnx/inputs.rs:150
↓ 4 callersFunctiontest_onnx_bin
(onnx: &[u8], bin: &[u8])
kn-cuda-eval/tests/root/runner.rs:109
↓ 4 callersMethodto_onnx_result
(self, path: impl AsRef<Path>)
kn-graph/src/onnx/result.rs:75
↓ 4 callersMethodtry_unwrap_fixed
(self)
kn-graph/src/shape.rs:278
↓ 4 callersFunctionunwrap_2
(slice: &[i64])
kn-graph/src/onnx/load.rs:1408
↓ 4 callersMethodvisit_scalable_value
(&mut self, value: Value)
kn-cuda-eval/src/planner.rs:590
↓ 3 callersFunction__float_simpl_cosf
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:3515
↓ 3 callersFunction__float_simpl_sinf
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:3507
↓ 3 callersFunction__floats2half2_rn
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1253
↓ 3 callersFunction__hmax_nan
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:3883
↓ 3 callersFunction__int2half_rn
defined(__CUDACC__) || defined(_NVHPC_CUDA) */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1609
↓ 3 callersFunction__internal_device_fast_bf16exp
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:3701
↓ 3 callersFunction__ll2half_rn
defined(__CUDACC__) || defined(_NVHPC_CUDA) */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1941
↓ 3 callersFunction__nv_bfloat16
kn-cuda-sys/doc_headers/cuda_include/cuda_bf16.hpp:253
↓ 3 callersFunction__uint2half_rn
defined(__CUDACC__) || defined(_NVHPC_CUDA) */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1741
↓ 3 callersFunction__ull2half_rn
defined(__CUDACC__) || defined(_NVHPC_CUDA) */
kn-cuda-sys/doc_headers/cuda_include/cuda_fp16.hpp:1873
↓ 3 callersMethodadd_named_axis
(&mut self, name: &str, value: Size)
kn-graph/src/onnx/loader.rs:98
↓ 3 callersMethodas_c_str
(self)
kn-graph/src/dtype.rs:150
↓ 3 callersMethodas_shape
(&self, graph: &Graph)
kn-graph/src/onnx/typed_value.rs:89
← previousnext →101–200 of 1,306, ranked by callers