The new cases added for quantized matmuls are:
1. vec-vec
2. vec-mat
3. mat-vec
each of which are now lowered to expand(s), quantized_matmul, and
collapse.
1. onnx.MatMulInteger now converts to aten.matmul instead of aten.mm
2. aten.matmul, for ranks >=2, now allows quantized inputs and will
lower to linalg::quantized_matmul or linalg::quantized_batch_matmul.
3. added AtenMatmulOp to the FuseQuantizeOps rewrite patters
QuantizeOperands, QuantizeTransposedOperands, and QuantizeAccumulator
4. added several tests, including some to test AtenMmOp with varying
quantization signed-ness.
5. a quantized matmul mat-vec test is added to verify the failure to
lower to linalg; cleaned of out-of-date code related to common
torch-mlir lowering xfails.
6. in debugging a real model with quantized matmuls, I found a bug on
the scalarize-shapes pass which resulted from the aten.full op folder
returning an incompatible result type. This is fixed by the small change
here to
[lib/Dialect/Torch/IR/TorchOps.cpp](https://github.com/llvm/torch-mlir/compare/main...zjgarvey:torch-mlir:MatMulIntegerFix?expand=1#diff-dc8ed165c207918e606490eee3984b1ad51d7034e6aac36fc046bf47f6f03f4f).
- Added linalg lowering for `AtenFloorDivideScalarOp`
- Needed `AtenDivScalarModeOp` for the decomp.
- Added linalg lowering for `AtenDivScalarModeOp`
- Moved linalg payload logic to `createDivModePayload()` since the logic
was nearly identical for both `AtenDivScalarModeOp` and
`AtenDivTensorModeOp`. Just a template function
- Added `AtenDivScalarModeOp` lowering for stablehlo
Pytorch's
[`torch.floor_divide()`](https://pytorch.org/docs/stable/generated/torch.floor_divide.html)
in a previous version (for a reason unknown to me) preformed a
truncation instead of "floor". The already implemented op
`AtenFloorDivideTensorOp` was done before this change. However, this
wasn't caught because our testcases only tested positive floor division.
I changed this to floor as well as adding a few test cases.
If there is only a single value scattered there can be an implicit batch
dimension. This includes a check for the implicit batch dimension when
reshaping the update tensor. It includes an e2e test to verify
correctness.
Fix the case PrimListUnpackOp's result num is not equal to PrimList
length.
See the following example:
```python
def forward(self, x):
if len(x.shape) == 5:
b0, t, c0, h0, w0 = x.shape
b, c, h, w = torch.mul(b0, t), c0, h0, w0
else:
b1, c1, h1, w1 = x.shape
b, c, h, w = b1, c1, h1, w1
res = torch.reshape(x, [b, c, h, w])
return res
```
Without this fix, the following error message will occur：
```
/root/torch-mlir/externals/llvm-project/mlir/lib/IR/PatternMatch.cpp:118: virtual void mlir::RewriterBase::replaceOp(mlir::Operation *, mlir::ValueRange): Assertion `op->getNumResults() == newValues.size() && "incorrect # of replacement values"' failed.
```
Previously, it could only handle the situations where outputsize == (1,
1) or outputsize == (input_H, input_W). Now it supports all situations
where input_H % output_H== 0 && input_W % output_W == 0
1. Changes the linalg lowering for dequantization ops to always sign
cast to float to prevent misrepresenting uint32 overflow on subtraction
with zero point.
2. Adds a basic quantized model test which only quantizes and
dequantizes and now passes with these changes in linalg and onnx
configs.
3. Changes the aten.mm lowering to allow mismatched quantized types.
4. If a quantized matmul arg is uint8, we shift by 128 to faithfully
represent the quantization as a signed i8 quantization. This worked fine
in the AtenMmOp lowering, but I'd be happy to move it to a rewrite in
FuseQuantizedOps.cpp instead if that seems more appropriate.
With the changes 3 and 4, the QuantizedMLP_basic and
QuantizedSingleLayer_basic e2e tests now passes with the onnx config.
They replaced all `c10::optional` usages with `std::optional` in
torchgen'd code in
fb90b4d4b2
causing the LTC build to break.
Replacing all usages of `c10::optional` with `std::optional` in
`projects/ltc` has fixed the issue
Issue: #3120
Added error message when adding new torch op to
[torch_ods_gen.py](https://github.com/llvm/torch-mlir/compare/main...IanWood1:torch-mlir:ods_gen_error_message?expand=1#diff-889b60b904ed67a5065a14e8de6fc89e00e199577e4d2bfa134ac4d1c89832d2).
New message displays which op key is failing and possible matches in the
torch `Registry`.
```Op does not match any Torch ops in Registry
Given op:
"aten::hardtanh_wrong : (Tensor, Scalar) -> (Tensor)"
Possible matches:
"aten::hardshrink : (Tensor, Scalar) -> (Tensor)"
"aten::hardtanh_ : (Tensor, Scalar, Scalar) -> (Tensor)"
"aten::hardtanh : (Tensor, Scalar, Scalar) -> (Tensor)"
"aten::clamp_min : (Tensor, Scalar) -> (Tensor)"
"aten::linalg_cond : (Tensor, Scalar?) -> (Tensor)"```
Also, ran black formatting on file. Based on LLVM style guides this seems to be correct, but I can revert the formatting if needed.
This PR only performs a lit test. In lieu of an e2e test, https://github.com/nod-ai/SHARK-TestSuite/pull/142 makede sure that the lowering works & the numbers check out.
Co-authored-by: Xida Ren <xida.ren.dev@gmail.com>
Shapes can be processed as tensors to represent the set of dimensions.
As reshapes take a list of scalars this can result in a single dynamic
dimension blocking the adjacent static dimensions.
This pass attempts to de-couple tensor computations related to shapes
and propagate values to better support lowering scalar tensor
computations.
There is an issue with stablehlo's linalg compilation. Canonicalization
appears to cleanup the issues until we can determine what in
mlir/stablehlo is the source of the issue.
See the related issues here:
[SHARK-Turbine#556](https://github.com/nod-ai/SHARK-Turbine/issues/556)
1. Adds uint8 casting to onnx.Cast op
2. Fixes an issue with onnx.DequantizeLinear when the scale comes with
shape [1].
3. Adds support for unsigned types in an AtenItemOp folder
4. Adds a simpler quantized model for easier debugging
5. Adds a fusion pass to convert [quant -> dequant -> transpose -> mm]
patterns to [transpose -> quant -> mm].
6. Moved some xfails that are still not passing, but for different
reasons than onnx.cast failures.
This was found while tracing backwards graphs: the convolution_backwards
op will return None if the first result is not needed. Confirmed by
defining a custom op with a `Tensor` return signature and having its
meta kernel return None.
Two e2e tests (AdaptiveAveragePool1/2dUnitOutputSizeDynamic) were
failing due to numerics. This was as a result of passing -1 as the
kernel size in the lowering for the corresponding onnx op
GlobalAveragePool.
Reshaping tensors depend on directly matching individual dimensions to
their corresponding dim in the `torch.view` reshape dimensions. This
involves decoupling dynamic dimensions from their static counterparts
and support cleanup / canonicalization.
This commit adds the OnnxToTorch lowering for the Mish, Softplus,
HardSwish, Trilu, ThresholdedRelu op
Signed-Off By: Vivek Khandelwal <vivekkhandelwal1424@gmail.com>
The previous conversions for AtenAdaptiveAvgPool1dOp and
AtenAdaptiveMaxPool2dOp are refactored into a general templated
conversion that works for all of the AtenAdaptive...PoolNdOp's.
New support is added for the following ops:
1. AtenAdaptiveMaxPool1d
2. AtenAdaptiveMaxPool3d
3. AtenAdaptiveAvgPool3d
Support is also provided for passing inputs without batch dimensions.
For example, applying adaptive_avg_pool2d to an input tensor of rank 3.
After [pytorch #118162](https://github.com/pytorch/pytorch/pull/118162)
gets down to torch-mlir, I'll add a test for AdaptiveMaxPool1d with
return_indices (which will pass with that upstream fix).
---------
Co-authored-by: James Newling <james.newling@gmail.com>
This adds support for converting DynamicQuantizeLinear from torch-onnx
to torch.
I could not get an e2e test to pass, since there seems to be some issues
with uint8 casting somewhere lower in the pipeline. For example
compiling with IREE for llvm-cpu, I would get either the correct zero
point (if zp < 128) or the correct zero-point minus 256 (if zp >= 128).
The output tensor seems to always return a tensor of zeros, which also
occurs when running uint8 examples through QuantizeLinear.
Edit: the first problem can be resolved by casting the output back to
uint8 on output, the second problem is resolved with PR #3018
Added support for dynamic shapes in `flattenusingints` op in tosa
dialect. Due to this some Argmax tests pass
This PR fixes this issue https://github.com/llvm/torch-mlir/issues/3004
The following tests pass after this PR
```
1. "ArgmaxIntModule_basic"
2. "ArgmaxIntModule_multiple_maxs"
3. "ArgmaxModule_basic"
```
zjgarvey
Vinayak Dev
penguin_wwy
Xinyu Yang
IanWood1
Aart Bik
Xinan Jiang(姜曦楠)
Rob Suderman
Yuanqiang Liu
Jae Hoon (Antonio) Kim
Xida Ren (Cedar)
Vivek Khandelwal
ptrifunovic98
Stella Laurenzo
schnkmwt
Xida Ren (Cedar)
Abhishek-TyRnT