llvm
/
torch-mlir
mirror of https://github.com/llvm/torch-mlir
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity
torch-mlir/test/Dialect/Basicpy/ops.mlir

44 lines
1.7 KiB
MLIR
Raw Normal View History

Add remaining pieces to capture full example models. * Adds Basicpy List, Tuple, Dict types and plumbs through C API. * Started debugging the issues around aten::conv2d capture, but a PyTorch bug is suspected. * Was able to manually verify that the basic conv2d forward test captures correctly with a workaround. * Need to resolve some printing issues upstream and move these tests to an integration test target (they take ~seconds to run).
2020-10-17 08:38:07 +08:00
// RUN: npcomp-opt -split-input-file %s | npcomp-opt | FileCheck --dump-input=fail %s
// -----
// CHECK-LABEL: @build_dict_generic
func @build_dict_generic() -> !basicpy.DictType {
// CHECK: basicpy.build_dict : () -> !basicpy.DictType
%0 = basicpy.build_dict : () -> !basicpy.DictType
return %0 : !basicpy.DictType
}
// -----
// CHECK-LABEL: @build_list_generic
func @build_list_generic(%arg0 : si32, %arg1 : si32) -> !basicpy.ListType {
// CHECK: basicpy.build_list %arg0, %arg1 : (si32, si32) -> !basicpy.ListType
%0 = basicpy.build_list %arg0, %arg1 : (si32, si32) -> !basicpy.ListType
return %0 : !basicpy.ListType
}
// -----
// CHECK-LABEL: @build_tuple_generic
func @build_tuple_generic(%arg0 : si32, %arg1 : si32) -> !basicpy.TupleType {
// CHECK: basicpy.build_tuple %arg0, %arg1 : (si32, si32) -> !basicpy.TupleType
%0 = basicpy.build_tuple %arg0, %arg1 : (si32, si32) -> !basicpy.TupleType
return %0 : !basicpy.TupleType
}
Add basicpy.numeric_constant op. * Going through TODOs on the PyTorch side, this is a big cause of them (not being able to have constants for signed/unsigned). * Added complex while in here since we're at the phase where it is better to just have things complete than partially done.
2020-11-24 11:20:26 +08:00
// -----
// CHECK-LABEL: @numeric_constant
func @numeric_constant() {
// CHECK: %num-1_si32 = basicpy.numeric_constant -1 : si32
%0 = basicpy.numeric_constant -1 : si32
// CHECK: %num1_ui32 = basicpy.numeric_constant 1 : ui32
%1 = basicpy.numeric_constant 1 : ui32
// CHECK: %num = basicpy.numeric_constant 2.000000e+00 : f32
%2 = basicpy.numeric_constant 2.0 : f32
// CHECK: %num_0 = basicpy.numeric_constant [2.000000e+00 : f32, 3.000000e+00 : f32] : complex<f32>
%3 = basicpy.numeric_constant [2.0 : f32, 3.0 : f32] : complex<f32>
Add initial TorchScript module importer It turns out that this was easiest to structure as a general IValue importer, since torch module are just one of the possible IValue's. We import the IValue object graph in a braindead fashion into basicpy ops and a new `torch.nn_module` op that is used to model the attributes/methods of a torch::jit::Module IValue. See `Torch/ops.mlir` for an example, and also check out the .py import tests in `frontends/pytorch/test/module_import`. As part of this change, a few housekeeping tasks: - extract some helpers from graph_importer.cpp - more helpers around the C API - misc touchups
2021-01-28 08:35:44 +08:00
// CHECK: %bool_true = basicpy.bool_constant true
%4 = basicpy.bool_constant true
// CHECK: %bool_false = basicpy.bool_constant false
%5 = basicpy.bool_constant false
Add basicpy.numeric_constant op. * Going through TODOs on the PyTorch side, this is a big cause of them (not being able to have constants for signed/unsigned). * Added complex while in here since we're at the phase where it is better to just have things complete than partially done.
2020-11-24 11:20:26 +08:00
return
}