tf-nightly-cpu 2.20.0.dev20250220__cp311-cp311-win_amd64.whl → 2.20.0.dev20250222__cp311-cp311-win_amd64.whl
Sign up to get free protection for your applications and to get access to all the features.
- tensorflow/_api/v2/compat/v1/summary/__init__.py +2 -2
- tensorflow/_api/v2/compat/v1/tpu/experimental/embedding/__init__.py +2 -2
- tensorflow/_api/v2/compat/v2/summary/__init__.py +10 -10
- tensorflow/_api/v2/compat/v2/summary/experimental/__init__.py +4 -4
- tensorflow/_api/v2/compat/v2/tpu/experimental/embedding/__init__.py +2 -2
- tensorflow/_api/v2/summary/__init__.py +10 -10
- tensorflow/_api/v2/summary/experimental/__init__.py +4 -4
- tensorflow/_api/v2/tpu/experimental/embedding/__init__.py +2 -2
- tensorflow/compiler/mlir/stablehlo/stablehlo_extension.pyd +0 -0
- tensorflow/compiler/tf2tensorrt/_pywrap_py_utils.pyd +0 -0
- tensorflow/compiler/tf2xla/ops/_xla_ops.so +0 -0
- tensorflow/include/external/llvm-project/mlir/include/mlir/Analysis/DataFlow/IntegerRangeAnalysis.h +12 -0
- tensorflow/include/external/llvm-project/mlir/include/mlir/Dialect/Math/IR/MathOps.h.inc +4 -0
- tensorflow/include/external/shardy/shardy/dialect/sdy/transforms/propagation/aggressive_factor_propagation.h +9 -0
- tensorflow/include/external/stablehlo/_virtual_includes/stablehlo_pass_utils/stablehlo/transforms/PassUtils.h +7 -0
- tensorflow/include/external/stablehlo/_virtual_includes/stablehlo_passes/stablehlo/transforms/PassUtils.h +7 -0
- tensorflow/include/external/stablehlo/_virtual_includes/version/stablehlo/dialect/Version.h +1 -1
- tensorflow/include/external/stablehlo/stablehlo/dialect/Version.h +1 -1
- tensorflow/include/external/stablehlo/stablehlo/transforms/PassUtils.h +7 -0
- tensorflow/include/tensorflow/compiler/xla/backends/cpu/codegen/kernel_api_ir_builder.h +3 -2
- tensorflow/include/tensorflow/compiler/xla/backends/cpu/runtime/convolution_thunk_internal.h +8 -10
- tensorflow/include/tensorflow/compiler/xla/backends/cpu/runtime/kernel_thunk.h +9 -3
- tensorflow/include/tensorflow/compiler/xla/backends/cpu/runtime/work_queue.h +81 -19
- tensorflow/include/tensorflow/compiler/xla/codegen/kernel_spec.h +24 -7
- tensorflow/include/tensorflow/compiler/xla/hlo/ir/hlo_casting_utils.h +0 -44
- tensorflow/include/tensorflow/compiler/xla/hlo/ir/hlo_instruction.h +12 -0
- tensorflow/include/tensorflow/compiler/xla/mlir_hlo/_virtual_includes/stablehlo_extension_pass_inc_gen/stablehlo_ext/transforms/passes.h.inc +149 -4
- tensorflow/include/tensorflow/compiler/xla/mlir_hlo/stablehlo_ext/transforms/passes.h.inc +149 -4
- tensorflow/include/tensorflow/compiler/xla/pjrt/distributed/client.h +5 -0
- tensorflow/include/tensorflow/compiler/xla/pjrt/gpu/se_gpu_pjrt_client.h +1 -92
- tensorflow/include/tensorflow/compiler/xla/pjrt/gpu/se_gpu_topology_description.h +126 -0
- tensorflow/include/tensorflow/compiler/xla/pjrt/pjrt_stream_executor_client.h +1 -49
- tensorflow/include/tensorflow/compiler/xla/pjrt/pjrt_stream_executor_device_description.h +75 -0
- tensorflow/include/tensorflow/compiler/xla/pjrt/plugin/xla_cpu/cpu_execute_options.h +57 -0
- tensorflow/include/tensorflow/compiler/xla/pjrt/plugin/xla_cpu/cpu_topology.h +4 -0
- tensorflow/include/tensorflow/compiler/xla/service/constant_value.h +1 -0
- tensorflow/include/tensorflow/compiler/xla/service/hlo_module_util.h +52 -1
- tensorflow/include/tensorflow/compiler/xla/service/hlo_proto_util.h +0 -12
- tensorflow/include/tensorflow/compiler/xla/tsl/concurrency/async_value.h +50 -21
- tensorflow/include/tensorflow/compiler/xla/tsl/framework/convolution/eigen_spatial_convolutions-inl.h +5 -5
- tensorflow/include/tensorflow/core/kernels/data/experimental/random_access_ops.h +0 -2
- tensorflow/include/tensorflow/core/kernels/eigen_attention.h +4 -4
- tensorflow/include/tensorflow/core/kernels/eigen_backward_cuboid_convolutions.h +6 -6
- tensorflow/include/tensorflow/core/kernels/eigen_backward_spatial_convolutions.h +10 -8
- tensorflow/include/tensorflow/core/kernels/eigen_cuboid_convolution.h +6 -6
- tensorflow/include/tensorflow/core/kernels/eigen_pooling.h +12 -12
- tensorflow/include/tensorflow/core/public/release_version.h +39 -0
- tensorflow/include/tensorflow/core/public/version.h +112 -127
- tensorflow/include/tensorflow/python/eager/pywrap_tfe.h +1 -1
- tensorflow/include/xla/backends/cpu/codegen/kernel_api_ir_builder.h +3 -2
- tensorflow/include/xla/backends/cpu/runtime/convolution_thunk_internal.h +8 -10
- tensorflow/include/xla/backends/cpu/runtime/kernel_thunk.h +9 -3
- tensorflow/include/xla/backends/cpu/runtime/work_queue.h +81 -19
- tensorflow/include/xla/codegen/kernel_spec.h +24 -7
- tensorflow/include/xla/hlo/ir/hlo_casting_utils.h +0 -44
- tensorflow/include/xla/hlo/ir/hlo_instruction.h +12 -0
- tensorflow/include/xla/mlir_hlo/_virtual_includes/stablehlo_extension_pass_inc_gen/stablehlo_ext/transforms/passes.h.inc +149 -4
- tensorflow/include/xla/mlir_hlo/stablehlo_ext/transforms/passes.h.inc +149 -4
- tensorflow/include/xla/pjrt/distributed/client.h +5 -0
- tensorflow/include/xla/pjrt/gpu/se_gpu_pjrt_client.h +1 -92
- tensorflow/include/xla/pjrt/gpu/se_gpu_topology_description.h +126 -0
- tensorflow/include/xla/pjrt/pjrt_stream_executor_client.h +1 -49
- tensorflow/include/xla/pjrt/pjrt_stream_executor_device_description.h +75 -0
- tensorflow/include/xla/pjrt/plugin/xla_cpu/cpu_execute_options.h +57 -0
- tensorflow/include/xla/pjrt/plugin/xla_cpu/cpu_topology.h +4 -0
- tensorflow/include/xla/service/constant_value.h +1 -0
- tensorflow/include/xla/service/hlo_module_util.h +52 -1
- tensorflow/include/xla/service/hlo_proto_util.h +0 -12
- tensorflow/include/xla/tsl/concurrency/async_value.h +50 -21
- tensorflow/include/xla/tsl/framework/convolution/eigen_spatial_convolutions-inl.h +5 -5
- tensorflow/lite/experimental/microfrontend/python/ops/_audio_microfrontend_op.so +0 -0
- tensorflow/lite/python/analyzer_wrapper/_pywrap_analyzer_wrapper.pyd +0 -0
- tensorflow/lite/python/interpreter_wrapper/_pywrap_tensorflow_interpreter_wrapper.pyd +0 -0
- tensorflow/lite/python/optimize/_pywrap_tensorflow_lite_calibration_wrapper.pyd +0 -0
- tensorflow/python/_pywrap_dtensor_device.pyd +0 -0
- tensorflow/python/_pywrap_mlir.pyd +0 -0
- tensorflow/python/_pywrap_parallel_device.pyd +0 -0
- tensorflow/python/_pywrap_quantize_training.pyd +0 -0
- tensorflow/python/_pywrap_tensorflow_internal.pyd +0 -0
- tensorflow/python/_pywrap_tfcompile.pyd +0 -0
- tensorflow/python/_pywrap_tfe.pyd +0 -0
- tensorflow/python/client/_pywrap_debug_events_writer.pyd +0 -0
- tensorflow/python/client/_pywrap_device_lib.pyd +0 -0
- tensorflow/python/client/_pywrap_events_writer.pyd +0 -0
- tensorflow/python/client/_pywrap_tf_session.pyd +0 -0
- tensorflow/python/compat/compat.py +1 -1
- tensorflow/python/data/experimental/service/_pywrap_server_lib.pyd +0 -0
- tensorflow/python/data/experimental/service/_pywrap_utils_exp.pyd +0 -0
- tensorflow/python/eager/imperative_grad.py +5 -5
- tensorflow/python/eager/polymorphic_function/atomic_function.py +1 -1
- tensorflow/python/eager/polymorphic_function/compiler_ir.py +1 -1
- tensorflow/python/eager/polymorphic_function/polymorphic_function.py +45 -41
- tensorflow/python/eager/tape.py +2 -2
- tensorflow/python/framework/_dtypes.pyd +0 -0
- tensorflow/python/framework/_op_def_library_pybind.pyd +0 -0
- tensorflow/python/framework/_op_def_registry.pyd +0 -0
- tensorflow/python/framework/_proto_comparators.pyd +0 -0
- tensorflow/python/framework/_pywrap_python_op_gen.pyd +0 -0
- tensorflow/python/framework/_test_metrics_util.pyd +0 -0
- tensorflow/python/grappler/_pywrap_tf_cluster.pyd +0 -0
- tensorflow/python/grappler/_pywrap_tf_item.pyd +0 -0
- tensorflow/python/grappler/_pywrap_tf_optimizer.pyd +0 -0
- tensorflow/python/lib/core/_pywrap_py_func.pyd +0 -0
- tensorflow/python/lib/io/_pywrap_file_io.pyd +0 -0
- tensorflow/python/lib/io/_pywrap_record_io.pyd +0 -0
- tensorflow/python/ops/summary_ops_v2.py +5 -1
- tensorflow/python/platform/_pywrap_tf2.pyd +0 -0
- tensorflow/python/profiler/internal/_pywrap_profiler.pyd +0 -0
- tensorflow/python/profiler/internal/_pywrap_profiler_plugin.pyd +0 -0
- tensorflow/python/saved_model/pywrap_saved_model.pyd +0 -0
- tensorflow/python/tpu/_pywrap_sparse_core_layout.pyd +0 -0
- tensorflow/python/tpu/_pywrap_tpu_embedding.pyd +0 -0
- tensorflow/python/tpu/tpu_embedding_v3.py +14 -7
- tensorflow/python/tpu/tpu_embedding_v3_checkpoint_adapter.py +10 -1
- tensorflow/python/util/_pywrap_checkpoint_reader.pyd +0 -0
- tensorflow/python/util/_pywrap_kernel_registry.pyd +0 -0
- tensorflow/python/util/_pywrap_stat_summarizer.pyd +0 -0
- tensorflow/python/util/_pywrap_tfprof.pyd +0 -0
- tensorflow/python/util/_pywrap_transform_graph.pyd +0 -0
- tensorflow/python/util/_pywrap_utils.pyd +0 -0
- tensorflow/python/util/_tf_stack.pyd +0 -0
- tensorflow/tools/pip_package/setup.py +2 -2
- tensorflow/xla_aot_runtime_src/xla/tsl/concurrency/async_value.cc +26 -51
- {tf_nightly_cpu-2.20.0.dev20250220.dist-info → tf_nightly_cpu-2.20.0.dev20250222.dist-info}/METADATA +1 -1
- {tf_nightly_cpu-2.20.0.dev20250220.dist-info → tf_nightly_cpu-2.20.0.dev20250222.dist-info}/RECORD +128 -123
- tensorflow/include/tensorflow/compiler/xla/backends/cpu/runtime/concurrency.h +0 -77
- tensorflow/include/xla/backends/cpu/runtime/concurrency.h +0 -77
- {tf_nightly_cpu-2.20.0.dev20250220.dist-info → tf_nightly_cpu-2.20.0.dev20250222.dist-info}/WHEEL +0 -0
- {tf_nightly_cpu-2.20.0.dev20250220.dist-info → tf_nightly_cpu-2.20.0.dev20250222.dist-info}/entry_points.txt +0 -0
- {tf_nightly_cpu-2.20.0.dev20250220.dist-info → tf_nightly_cpu-2.20.0.dev20250222.dist-info}/top_level.txt +0 -0
@@ -1,77 +0,0 @@
|
|
1
|
-
/* Copyright 2024 The OpenXLA Authors.
|
2
|
-
|
3
|
-
Licensed under the Apache License, Version 2.0 (the "License");
|
4
|
-
you may not use this file except in compliance with the License.
|
5
|
-
You may obtain a copy of the License at
|
6
|
-
|
7
|
-
http://www.apache.org/licenses/LICENSE-2.0
|
8
|
-
|
9
|
-
Unless required by applicable law or agreed to in writing, software
|
10
|
-
distributed under the License is distributed on an "AS IS" BASIS,
|
11
|
-
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
12
|
-
See the License for the specific language governing permissions and
|
13
|
-
limitations under the License.
|
14
|
-
==============================================================================*/
|
15
|
-
|
16
|
-
#ifndef XLA_BACKENDS_CPU_RUNTIME_CONCURRENCY_H_
|
17
|
-
#define XLA_BACKENDS_CPU_RUNTIME_CONCURRENCY_H_
|
18
|
-
|
19
|
-
#include <cstdint>
|
20
|
-
#include <functional>
|
21
|
-
#include <memory>
|
22
|
-
#include <type_traits>
|
23
|
-
|
24
|
-
#include "tsl/platform/logging.h"
|
25
|
-
|
26
|
-
#define EIGEN_USE_THREADS
|
27
|
-
#include "unsupported/Eigen/CXX11/Tensor"
|
28
|
-
#include "unsupported/Eigen/CXX11/ThreadPool"
|
29
|
-
|
30
|
-
namespace xla::cpu {
|
31
|
-
|
32
|
-
// A collection of concurrency primitives for use in the XLA CPU runtime.
|
33
|
-
|
34
|
-
// Schedules `n` tasks on the `intra_op_threadpool`, calling `F` for each index
|
35
|
-
// in the [0, n) range. Returns immediately after scheduling all tasks. It's a
|
36
|
-
// caller's responsibility to wait for all tasks to finish.
|
37
|
-
template <typename F,
|
38
|
-
std::enable_if_t<std::is_invocable_v<F, int64_t>>* = nullptr>
|
39
|
-
void ScheduleAll(const Eigen::ThreadPoolDevice* intra_op_threadpool, int64_t n,
|
40
|
-
F&& f) {
|
41
|
-
DCHECK(n >= 0) << "n must be non-negative";
|
42
|
-
|
43
|
-
// Short-circuit the case of no tasks.
|
44
|
-
if (n == 0) return;
|
45
|
-
|
46
|
-
// Short-circuit the case of a single task.
|
47
|
-
if (n == 1) {
|
48
|
-
f(0);
|
49
|
-
return;
|
50
|
-
}
|
51
|
-
|
52
|
-
// Heap-allocated state that manages concurrent execution of `f`.
|
53
|
-
struct State {
|
54
|
-
State(const Eigen::ThreadPoolDevice* intra_op_threadpool, F&& f)
|
55
|
-
: intra_op_threadpool(intra_op_threadpool), f(std::forward<F>(f)) {}
|
56
|
-
|
57
|
-
void Execute(std::shared_ptr<State> self, int64_t start, int64_t end) {
|
58
|
-
while (end - start > 1) {
|
59
|
-
uint64_t mid = (start + end) / 2;
|
60
|
-
intra_op_threadpool->getPool()->Schedule(
|
61
|
-
std::bind(&State::Execute, this, self, mid, end));
|
62
|
-
end = mid;
|
63
|
-
}
|
64
|
-
f(start);
|
65
|
-
}
|
66
|
-
|
67
|
-
const Eigen::ThreadPoolDevice* intra_op_threadpool;
|
68
|
-
F f;
|
69
|
-
};
|
70
|
-
|
71
|
-
auto s = std::make_shared<State>(intra_op_threadpool, std::forward<F>(f));
|
72
|
-
s->Execute(std::move(s), 0, n);
|
73
|
-
}
|
74
|
-
|
75
|
-
} // namespace xla::cpu
|
76
|
-
|
77
|
-
#endif // XLA_BACKENDS_CPU_RUNTIME_CONCURRENCY_H_
|
@@ -1,77 +0,0 @@
|
|
1
|
-
/* Copyright 2024 The OpenXLA Authors.
|
2
|
-
|
3
|
-
Licensed under the Apache License, Version 2.0 (the "License");
|
4
|
-
you may not use this file except in compliance with the License.
|
5
|
-
You may obtain a copy of the License at
|
6
|
-
|
7
|
-
http://www.apache.org/licenses/LICENSE-2.0
|
8
|
-
|
9
|
-
Unless required by applicable law or agreed to in writing, software
|
10
|
-
distributed under the License is distributed on an "AS IS" BASIS,
|
11
|
-
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
12
|
-
See the License for the specific language governing permissions and
|
13
|
-
limitations under the License.
|
14
|
-
==============================================================================*/
|
15
|
-
|
16
|
-
#ifndef XLA_BACKENDS_CPU_RUNTIME_CONCURRENCY_H_
|
17
|
-
#define XLA_BACKENDS_CPU_RUNTIME_CONCURRENCY_H_
|
18
|
-
|
19
|
-
#include <cstdint>
|
20
|
-
#include <functional>
|
21
|
-
#include <memory>
|
22
|
-
#include <type_traits>
|
23
|
-
|
24
|
-
#include "tsl/platform/logging.h"
|
25
|
-
|
26
|
-
#define EIGEN_USE_THREADS
|
27
|
-
#include "unsupported/Eigen/CXX11/Tensor"
|
28
|
-
#include "unsupported/Eigen/CXX11/ThreadPool"
|
29
|
-
|
30
|
-
namespace xla::cpu {
|
31
|
-
|
32
|
-
// A collection of concurrency primitives for use in the XLA CPU runtime.
|
33
|
-
|
34
|
-
// Schedules `n` tasks on the `intra_op_threadpool`, calling `F` for each index
|
35
|
-
// in the [0, n) range. Returns immediately after scheduling all tasks. It's a
|
36
|
-
// caller's responsibility to wait for all tasks to finish.
|
37
|
-
template <typename F,
|
38
|
-
std::enable_if_t<std::is_invocable_v<F, int64_t>>* = nullptr>
|
39
|
-
void ScheduleAll(const Eigen::ThreadPoolDevice* intra_op_threadpool, int64_t n,
|
40
|
-
F&& f) {
|
41
|
-
DCHECK(n >= 0) << "n must be non-negative";
|
42
|
-
|
43
|
-
// Short-circuit the case of no tasks.
|
44
|
-
if (n == 0) return;
|
45
|
-
|
46
|
-
// Short-circuit the case of a single task.
|
47
|
-
if (n == 1) {
|
48
|
-
f(0);
|
49
|
-
return;
|
50
|
-
}
|
51
|
-
|
52
|
-
// Heap-allocated state that manages concurrent execution of `f`.
|
53
|
-
struct State {
|
54
|
-
State(const Eigen::ThreadPoolDevice* intra_op_threadpool, F&& f)
|
55
|
-
: intra_op_threadpool(intra_op_threadpool), f(std::forward<F>(f)) {}
|
56
|
-
|
57
|
-
void Execute(std::shared_ptr<State> self, int64_t start, int64_t end) {
|
58
|
-
while (end - start > 1) {
|
59
|
-
uint64_t mid = (start + end) / 2;
|
60
|
-
intra_op_threadpool->getPool()->Schedule(
|
61
|
-
std::bind(&State::Execute, this, self, mid, end));
|
62
|
-
end = mid;
|
63
|
-
}
|
64
|
-
f(start);
|
65
|
-
}
|
66
|
-
|
67
|
-
const Eigen::ThreadPoolDevice* intra_op_threadpool;
|
68
|
-
F f;
|
69
|
-
};
|
70
|
-
|
71
|
-
auto s = std::make_shared<State>(intra_op_threadpool, std::forward<F>(f));
|
72
|
-
s->Execute(std::move(s), 0, n);
|
73
|
-
}
|
74
|
-
|
75
|
-
} // namespace xla::cpu
|
76
|
-
|
77
|
-
#endif // XLA_BACKENDS_CPU_RUNTIME_CONCURRENCY_H_
|
{tf_nightly_cpu-2.20.0.dev20250220.dist-info → tf_nightly_cpu-2.20.0.dev20250222.dist-info}/WHEEL
RENAMED
File without changes
|
File without changes
|
File without changes
|