follow up #761:
This patch updates the `torch_mlir::convertTensorToMlirElementsAttr()`
method to enable the creation of tensors whose base type is Float16.
This patch also adds a test to validate the IR generation, and it
updates the test for importing tensors of various types.
PyTorch recently added support for `dim=None` in the `torch.var`
(5ca9b2b6fa)
and `torch.std`op (eb0e30e0bc).
This commit adds the corresponding support in torch-mlir.
Signed-Off By: Vivek Khandelwal<vivek@nod-labs.com>
In some cases, users know that a traced graph is valid for a wider set
of shapes than they originally traced it with. Provide an option for
users to ignore the shapes in the traced graph when they know it is
legal.
Fixes#997
* Replace CHECK_EQ with TORCH_CHECK_EQ
* Check value of TORCH_MLIR_USE_INSTALLED_PYTORCH during LTC build
* Update LTC XFAIL with NewZerosModule ops
* Explicitly blacklist _like ops
* Automatically blacklist new_/_like ops
* Prune away unused Python dependencies from LTC
* Add flag to disable LTC
* Autogen dummy _REFERENCE_LAZY_BACKEND library when LTC is disabled
* Implement compute_shape_var
* Removed Var tests from XFAIL Set
* XFAIL tests using _local_scalar_dense or index.Tensor
* Add StdDim tests to XFAIL set
* Autogen aten::cat
* Changed Example MLIR backend to Reference MLIR backend
* Moved reference_ltc_backend into csrc
* Merged sys_utils.h
* Renamed reference_ltc_backend to reference_lazy_backend
* Addressed review comments
* Update docs with new library name
* Removed _REFERENCE_LAZY_BACKEND from .gitignore
* Added reference_lazy_backend to the TorchMLIRPythonModules dependency list
Fixed typo in `ltc_examples.md`
Missed instance where `ltc_backend` was used instead of `lazy_backend`.
- Pruned number of xfailed e2e LTC tests from 305 to 134
- Reviewed every failure to ensure the error genuinely warrants an xfail
- Fixed bug where non-tensor outputs of LTC computation had `.to('cpu')` called, which caused a failure and inflated the xfail count
- Fixed bug with `HBC_basic` test where a constant tensor was created in its constructor without being declared as a buffer, which prevented the device from being updated when the parent `torch.nn.Module` got moved to the `lazy` device
- Note that this test is still xfail'd due to some unsupported ops. Left a comment about some potential issues that may arise if it gets reenabled in the future
- Updated autogen `GeneratedTorchOps.td` to reflect the latest set of supported ops
- Renamed `aten.zero.functionalization` to `aten.zero` to reflect upstream PyTorch changes
* Added e2e LTC Torch MLIR tests
* Fix seed for reproducability
* Check if computation is None before getting debug string
* Updated unit tests, and added numeric tests
* Print name of the model layer that fails numeric validation
* Run LTC e2e test with CI/CD
* Set seed in main function, instead of beginning of execution
* Add comment to specify number of digits of precision
* Fixed typo
* Remove tests for LTC example models
* Added LTC option to torchscript e2e
* Implement compile and run for LTC e2e test
* xfail all tests that use ops that aren't currently supported
* Update native function definitions
* Add ops to support bert lowering
- Add empty_strided and as_strided
- Restore zeros_like to op blacklist (Without this, tensors will be unintentionally created with a CPU device rather than lazy)
- Check for composite implicit ops and add device data IR
- Also fix codegen for functionalization
* Add autogen to CMakeList
* Remove PyTorch submodule
* Reduced BERT model size
* Print Mark Step status in Torch MLIR LTC debug string
* Apply fixes to work with latest upstream/main
- Pass importOptions into getMlirTypeFromTorchType during NodeImporter::importNode
Without this, the tensor type created may have a mismatched type as ImportOptions may cause vtensor to be used instead of tensor
* Update shape inference functions
- Fixed compute_shape_native_batch_norm when mean and var are uninitialized
Previously, the number of shapes returned would be <3 if either mean or val was didn't exist. Instead, we now initialize them with a vector matching the number of channels.
- Implemented compute_shape_mul
- Fixed bug in reshape shape inference error message
* Get MLIR backend more consistent with TS backend
- Remove LazyNativeFunctions::_unsafe_view from autogen
- Blacklist ops to make JIT graph more like output of TS backend
- Print graph when SSA value has mismatch of types and results
- Remove normalize_index from LazyShapeInference
- Fix seeds for LTC example models
* Update and clean up shape inference functions
- Prune shape inference functions
- Add shape inference function for GenerateSlice
- Add shape inference function for GenerateCopy
Co-authored-by: Henry Tu <henry.tu@cerebras.net>
* Assume zero rank tensors are scalar
* Run RefineTypes pass on JIT Graph
* Rollback assumption that zero rank tensors are scalar
* Set numSizes to -1 for non-ranked tensors
* Rename RefineTypes to RefineTupleTypes
* Save InputOutputAliases to TorchMlirComputation
* Implement GetResultShape for TorchMlirLoweringContext
* Use optional return type for GetResultShape
* Remove support for aten::detach
With this op enabled, tensors were being copied, which resulted in incorrect aliasing.
* Add newline before printing I/O alias mapping
* Changed printout to use "Input param" as label instead of "Input"
* Remote shape inference function for aten::detach
* Moved implementation of SetUpAlias to MlirLoweringContext
As part of this change, TorchMlirComputation has been moved to the end of mlir_lowering_context.h so that it can access some new structs in TorchMlirLoweringContext
* Use updated PyTorch API
* Remove GetResultShape
Complements this upstream PyTorch PR: pytorch/pytorch#75828
This PR adds support for mapping input and output tensors which alias each other. (e.g. maps input weight tensor in parameter to the same tensor in output after a training iteration)
MLIR:
func @graph(%arg0: !torch.vtensor<[1,5],f32>, %arg1: !torch.vtensor<[1],si64>, ..., %arg6: !torch.vtensor<[10,5],f32>, %arg7: !torch.vtensor<[10],f32>, ...) {
...
return %arg0, %arg1, %17, %23, ... : !torch.vtensor<[1,5],f32>, !torch.vtensor<[1],si64>, !torch.vtensor<[10,5],f32>, !torch.vtensor<[10],f32>, ...
}
Input/Output Alias Mapping:
Output: 0 -> Input: 0
Output: 1 -> Input: 1
Output: 2 -> Input: 6
Output: 3 -> Input: 7
The aten::detach op has also been disabled in this PR to fix the issue of tensors not aliasing properly due to copying.
* Added JIT to MLIR lowering
Lowering to JIT is performed in a way similar to how it's done in the TS LTC backend. After a jit::Graph is constructed, it gets converted to a jit::Function, which is fed into the existing utility to generate an MlirModule in torch-mlir.
* Renamed `csrc/backend` to `csrc/base_lazy_backend`
This commit fixes the shape calculation for:
1.) aten.mean.dim
2.) aten.var.dim
3.) aten.sum.dim_IntList op
Also, it fixes the lowering of `aten.mean.dim` and
`aten.sum.dim_IntList` for handling the cases of empty dim list.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com
- Includes a canonicalizer for `aten.add.t`needed for successfully lowering the shape function
- Only offers support for statically sized index tensors when there is more than one
- Dynamic shape support remains for single indexing tensors
This enables building Pytorch from source in the CI.
The build should mostly hit the ccache.
Release builds will follow once we have some runtime on the CI.
In the interest of merging upstream LLVM quickly, a previous patch
(7f08169) updated the torch-mlir build to register all dialects and
passes through Python bindings. This patch limits the dialects and
passes to only those that are used in torch-mlir.
Key to this change are the removal of
`MLIRPythonExtension.RegisterEverything` and the introduction of a new
Python module (`_mlir_libs/_site_initialize_0.py`), where we register
the dialects and passes used by torch-mlir.
- Supports cases where the view op expands and collapses dims
simulataneously. This does not handle the case where it is neither
expanding nor collapsing (e.g. [2, 3] -> [3, 2])
- Additionally fixes a previous bug with adding 1-sized dims on both
sides of a tensor with aten.view
This patch makes some rudimentary changes to torch-mlir's use of MLIR
Python bindings to work with the most recent LLVM code. We can perhaps
do better by being more selective in what we link against, instead of
using `MLIRPythonExtension.RegisterEverything`.
This commit adds the support for negative dim cases for `aten.cat`,
`aten.slice.Tensor` and `aten.slice_scatter` op.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
The original conversion pattern for `AtenBatchNormOp` required that
the input rank be greater than 2; however, the only
expectation in the conversion pattern and in Pytorch is that the input
rank is greater than 1, since the second dimension of the input must
match the size of the `weight`, `bias`, `runningMean`, and
`runningVar` inputs. This commit fixes the `inputRank` check.
This commit adds the decomposition for `aten.var.dim` op.
This commit also make changes in the decomposition for `aten.var` op.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
Remove all the libtorch downloads. If the user sets
-DTORCH_MLIR_USE_INSTALLED_PYTORCH=OFF then just build from src.
Doesn't change developer workflow since we still default to local
PyTorch versions.
TEST: Build and verify all tests (except one xfail quant) pass on linux
This commit does three things:
1. Reverts some of the shape lib changes merged in
https://github.com/llvm/torch-mlir/pull/844
2. Updates the signature of `aten.sum_dim_IntList` that was recently
updated in
23bdb570cf
3. Replaces `aten.zero.functional` with `aten.zero`, updated in 960758b0b7
`aten.select_scatter` op.
This commit adds:
1. Lowering of `aten.slice_scatter` op into `tensor.insert_slice`
op.
2. Decomposes the `aten.select_scatter` op into `aten.slice_scater`
op.
Signed-Off-By: Prateek Gupta <gprateek93@gmail.com>
Temporarily revert to using PyTorch binaries until source builds
are ready to land.
TORCH_MLIR_USE_INSTALLED_PYTORCH can be turned to OFF if you want
to link against libtorch and/or source builds.
On my local machine, `unzip` didn't exist (producing a "command not
found" error), but CMake ignored the error. Although the build did
succeed (because it found a previously-built version of libtorch), it
seems better to abort builds on such failures, so this patch checks the
return code of all external process invocations.
Along similar lines, this patch also updates the shell scripts in
`build_tools` to extensively use double-quoting to prevent unintentional
word splitting or globbing. Since some of the scripts execute `rm`
while using shell variables, this patch also adds the preamble `set -u`
to abort execution if an undefined variable is referenced, so that we
reduce the chances of executing `rm -rf /` if the path expression
happens to refer to an undefined variable.
Add an option to cache libtorch/ releases if you don't want to
download the latest. Add an option to enable source builds.
TESTS:
macOS: verify with / without cache downloads
verify source builds -- shared and static
Linux: Build Tests and Release builds
A previous fix to the handling of size-1 dims in
`aten.view` (https://github.com/llvm/torch-mlir/pull/962) resulted in
the wrong grouping of dimensions when size-1 dims where between two
dims of size greater than 1. This commit fixes that.
TorchScript nodes like `prim::Load` and `prim::Store` aren't supported
in torch-mlir because they can't be lowered to backends, but such nodes
can occur in the TorchScript IR.
This patch adds a rudimentary translation from such nodes to
corresponding ops in the Torch dialect. Since we expected such nodes to
go away during lowering because of the SymbolDCE pass, this patch does
not add code to lower these ops beyond the Torch dialect.
This commit lowers `aten.matmul` to `linalg.BatchMatmul` under the
following conditions:
1. The result of matrix multiplication must have batch dimensions,
i.e., rank greater than 2.
2. The resultant matrix must have at most 1 dynamic batch dimension.
It also handles broadcasting of batch dimensions when batch dimensions
of the matrices are broadcastable.
Signed-Off-by: Gaurav Shukla <gaurav@nod-labs.com>
This commit fixes the shape function for `index.Tensor`, adding
support for multiple index tensors and `None`s in the indices
list. This commit also adds support for input tensors of rank greater
than 1. The lowering for `index.Tensor` still has the the limitation
that only a single index tensor along the first dimension of the input
tensor is supported.
Prior to this patch, the torch dialect included `AtenTriuOp` for
computing the upper triangular part of the input matrix, but there was
no code for lowering the op to the linalg dialect.
This patch adds code to generate a `linalg.generic` operation that
compares indices (computed using `linalg.index`) to choose between zero
or the original value (using `arith.select`). The lowering fails if the
number of dimensions are less than two. This patch also adds a few
end-to-end tests.
The MacOS builders are having linking trouble with the extension library.
Until it's fixed, all support for op extensions is disabled. It should be
easy to restore once the issue is resolved.
The function `AffineMap::inferFromExprList` does not work if the first
vector of expressions is empty, because it uses these expressions to
obtain the context. This prevented `aten.permute` from working for
inputs of 0-rank. This commit adds support for 0-rank inputs.
PyTorch allows new operators to be registered dynamically in modules.
Torch-mlir already makes it fairly straightforward to add support for
new operators, and this commit just extends that support to allow new
PyTorch ops to come from a external module.
This does *not* allow ops to be dynamically loaded into torch-mlir.
Torch-mlir must still be compiled with support built-in.
Add a `_torch_mlir_custom_op_example` subpackage to `torch_mlir` which
registers an demonstration op. It will not be imported by default when
importing torch_mlir. It's strictly for testing and documentation.
Adds an end-to-end test for the `torch_mlir_custom_op_example::identity` op.
With all these changes, we should now be actively testing PyTorch extension
support with all future patches.
Now that upstream exposes them nicely, we can use them.
I noticed that we had added stuff into the upstream_shape_helpers.py
file (which was supposed to stay pristine), so some more shape functions
need to be upstreamed.
Going forward, all shape functions should be upstreamed similar to
https://github.com/pytorch/pytorch/pull/76889 instead of added in this
file.
This commit adds lowering of `aten.div.Tensor_mode` op.
This commit also fixes formatting for the test file elementwise.py.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
This commit decomposes `aten.baddbmm` op into `aten.bmm`,
`aten.mul.Scalar`, and `aten.add.Tensor` op.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
use_tracing=True was behaving unexpectedly because the handling of
single arguments was happening after the torch.jit.trace call.
This also fixes the check to specifically test for a torch.Tensor or
TensorPlaceholder so that both lists and tuples would be correctly
handled.
This commit adds the decomposition of `aten.adaptive_avg_pool2d` op into
`aten.avg_pool2d` op. The current decomposition only supports cases where
input size is equal to the output size.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
This patch adds support for the torch.linalg.vector_norm op to the torch
dialect, including the necessary shape function. It also extends the
conversion of reduction operators to support lowering of
AtenLinalgVectorNormOp, in addition to adding a handful of end-to-end
tests to validate the lowering.
There exist several opportunities to make this lowering optimal and
robust. For instance, in its current form, the translation does not
support ord = 0, +inf, or -inf. For L1 norms, we don't need to raise
each element to the power 1.0. Similarly, L2 norms could benefit from
strength reduction. Since the canonicalization pass is not able to
apply these optimizations, we should consider applying them during the
linalg lowering itself.
We do this by inroducing a TensorPlaceholder class, which can be used to
specify dynamic sizes. Internally, we canonicalize all example inputs
to TensorPlaceholder's.
This commit also adds some basic testing, which was missing before.
In addition to updating the llvm-project submodule, this patch also:
1. updates shape functions and tests so that `func` and `call`
operations refer to the `func` dialect
2. avoid duplicate registration of dialects
The op `aten.rand_like` was missing a shape function, unit tests, and
the `dtype` argument was being ignored in its decomposition. This
commit fixes all three things.
A user might want to avoid the extra layer of multiprocessing libary for
debugging purpose. In such cases, the -s flag can be used to force
sequential execution.
This commit adds support for aten.max_pool2d, aten.max_pool2d_with_indices,
and aten.avg_pool2d op for the cases where ceil_mode = true.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
The preserve memory specifies that `If any of the input tensors is in channels_last format,
operator output should be in channels_last format` and hence can be
added as is in aten_empty_like op.
Fix the type promotion code for scalar only operation to return
TorchType which is the type tracked in ValueKnowledge.scalarType.
- Fix `getPromotedResultScalarType` to return Torch type.
- Add `getBuiltInTypeForTorchScalar` helper to convert scalar type
to builtin type before passing to the next level type promotion
helper `updateResultTypeState`.
- Add `setScalarType` helper to make setting ValueKnowledge.scalarType
easier.
This commit adds lowering of `aten.ge.float`, `aten.ge.float_int`,
`aten.ne.float_int`, `aten.gt.float_int` and `aten.ceil.float` op.
This commit also fixes formatting for the file scalar.py and scalar_comparison.py.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
The main changes are:
- Added `ValueKnowledge.scalarType` to track scalar type information.
- Added `ValueKnowledge.kind` to indicate the value kind.
- Modified the meet and join helper functions. The ValueKnowledge has
slightly more complicated state now so the meet and join function need
to look at the `kind` field in addition to just the type field.
- This commit adds support for `aten.mean.dim` op.
- It also adds a new test script `stats.py` for statistics related ops.
Signed-Off-by: Gaurav Shukla <gaurav@nod-labs.com>
This also has a fix for the adjustment of types of TupleConstruct
inputs, which I found when using this new functionality on a model.
Some scenarios in tracing create situations where the output of
TupleConstruct has a more refined type than the inputs.
This introduces a helper `adjustStaticInformationForValues` which
subsumes the `derefineValues` helper and the tensor static information
adjustment we were doing.
This commit decomposes `aten.to.dtype_layout` op into `aten.to.dtype` op.
This commit also fixes the formatting for the file type_conversion.py.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
This commit adds lowering of `aten.masked_fill.Scalar` op.
This commit also fixes the formatting of the file constant_alloc.py.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
This commit fixes the `ConstantPad2dStaticModule` test case by adding
the lowering of `aten.pad` operation. Previously the test case
mapped to `aten.constant_pad_nd` operation.
The `aten.pad` now decomposes into `aten.constant_pad_nd` operation.
Signed-Off-By: Prateek Gupta <prateek@nod-labs.com>
This patch updates the `torch_mlir::convertTensorToMlirElementsAttr()`
method to enable the creation of tensors whose base type is BFloat16.
This patch also adds a test to validate the IR generation, and it
updates the test for importing tensors of various types.
1. This commit adds lowering of "while-like" prim loop to scf.while
operation.
2. Adds lowering of "for-like" prim loops to scf.for operation.
Signed-Off-By: Prateek Gupta <prateek@nod-labs.com>
Fix the inplace update tensor issue we had
where the torchscript execution would update the input value inplace
resulting the actual test not being able to see the original input
value.
This commit adds more test cases `aten::index_put` op.
This commit also fixes formatting issues with the test file index_put.py
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
This commit adds lowering of `aten.ceil.float` op.
This commit also fixes formatting for the file scalar.py.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
Compiling torch-mlir against a source version of PyTorch or an official
wheel compiled with the new C++ stdlib ABI fails, as torch-mlir doesn't
know how to set compiler flags to remain compatible. This changes the
way torch-mlir looks at PyTorch and tries to more closely match the ABI
settings, regardless of whether it's the common official wheel or some
other version.
I wasn't able to find exactly what frontend situation created it, but
`torch.jit.trace` will sometimes create functions where the
`jit::Block`'s param node has refined tensor types. So we need to adjust
the function's formal param types to those refined types.
The updated LLVM code includes a patch to create bfloat16 array
attributes, thus enabling a different patch to torch-mlir to flesh out
support for the bfloat16 type.
* Add oneshot release snapshot for test/ondemand
Add some build scripts to test new release flow based on IREE.
Wont affect current builds, once this works well we can plumb it
in.
Build with manylinux docker
* Fixes a few issues found when debugging powderluv's setup.
* It is optional to link against Python3_LIBRARIES. Check that and don't do it if they don't exist for this config.
* Clean and auditwheel need to operate on sanitized package names. So "torch_mlir" vs "torch-mlir".
* Adds a pyproject.toml file that pins the build dependencies needed to detect both Torch and Python (the MLIR Python build was failing to detect because Numpy wasn't in the pip venv).
* Commented out auditwheel: These wheels are not PyPi compliant since they weak link to libtorch at runtime. However, they should be fine to deploy to users.
* Adds the --extra-index-url to the pip wheel command, allowing PyTorch to be found.
* Hack setup.py to remove the _mlir_libs dir before building. This keeps back-to-back versions from accumulating in the wheels for subsequent versions. IREE has a more principled way of doing this, but what I have here should work.
Co-authored-by: Stella Laurenzo <stellaraccident@gmail.com>
Added the dynamic registration of return function to the execution
engine. This makes sure that different/multiple return types are supported.
Also, updated the .style.yapf indentation to 4.
This makes it much easier to convert models and hides all the
ClassAnnotator complexity.
This also adds a new example `torchscript_resnet18_all_output_types.py`
which shows the ResNet18 IR for all output types.
Also,
- This moves `run_pipeline_with_repro_report` to
`torch_mlir.compiler_utils`.
That way, downstreams don't have to duplicate this list.
Also, remove "external config" feature, since it is subsumed by just
importing the test suite.
* shape: add shape transfer function for aten.neg
Prior to this patch, the list of shape transfer functions did not
include `aten.neg`, which resulted in errors like below.
```
error: unsupported by backend lowering: tensor with unknown rank or dtype
note: see current operation: %0 = "torch.aten.neg"(%arg0) :
(!torch.vtensor<[256,256],f32>) -> !torch.vtensor<*,f32>
note: this is likely due to a missing shape transfer function in shape_lib_gen.py
```
This patch fixes the problem by adding a shape transfer function to
reflect the point-wise nature of this operation.
* linalg: add translation of aten.neg operation
This patch adds a translation rule to lower `aten.neg` operations on
tensors to an `arith.negf` operation wrapped inside a `linalg.generic`
operation. This patch also adds a rudimentary test.
This commit adds lowering of `aten::max_pool2d_with_indices_backward` op.
This commit also fixes formatting issues in basic.py.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
NB: `shouldnt_normalize2` and `shouldnt_normalize3` currently XPASS i.e., args *will* successfully normalize despite being incorrect due to an [upstream bug](https://github.com/pytorch/pytorch/issues/75342).
This commit adds the following support to the op `nll_loss_backward`:
- `input` tensor can be rank-1
- `weight` parameter
- `reduction` parameter
- `target`, `grad_output`, `total_weight` can be rank-0
- Checks that input tensors are of the expected type
This commit adds support for multi-dimensional tensors as input to the
`_index_put_impl_` op. The support was to some degree already there,
since `ScatterOp` already supports multi-dimensional tensors. This
commit also adds a bit more error checking to `index_put` and
refactors the code for creating `ScatterOp`s to mimic the way one
would make a `Linalg::GenericOp`.
The issue was in the canonicalizer for torch.aten.ge.int -- in cases
where the operands were swapped, it would miscompile. This issue is
fixed and folding support generalized to `torch.aten.size.int < 0` as
well.
Fixes#716
This commit decomposes different variants of `aten.where.*` op into
`aten.where.Self` op. It covers `aten.where.Scalar`,
`aten.where.ScalarSelf` and `aten.where.ScalarOther` ops.
Signed-Off-by: Gaurav Shukla <gaurav@nod-labs.com>
This commit decomposes `aten.new_empty` op into `aten.empty.memory_format` op.
This commit also made a dtype fix to the constant tensor allocation like ops.
Earlier the dtype for the result was inferred from the result type; now, it's
being evaluated as per the original definition of the op.
Signed-Off By: Vivek Khandelwal <vivek@nod-labs.com>
Henry Tu
Tanyo Kwok
Sean Silva
Vivek Khandelwal
gpetters94
Ramiro Leal-Cavazos
Quinn Dawkins
Vidush Singhal
Alec
Jae Hoon (Antonio) Kim
Antonio Kim
PhaneeshB
Kevin Kiningham
powderluv
Ashay Rane
Ziheng Jiang
Kevin Kiningham
Maksim Levental
Prateek Gupta
Quinn Dawkins
Suraj Sudhir
Gaurav Shukla
erman-gurses
Albert Sandru
Bob Adolf
Yi Zhang
Prashant Kumar
yuhao