GitLab

Sigmoid approximation for GeLU was introduced in #1299 for Fp16. The sigmoid approximation is known to get better perf but lower accuracy. https://arxiv.org/pdf/1606.08415.pdf

Changed the doc for find_permutation(shape) to be more clear that it is finding the permutation that would make the shape standard

By converting to fp32 : fp16 3d-unet model accuracy comes out the same as FP32 accuracy.

By using reduce_sum method on Fp16 : accuracy comes out ~0.9% lower compared to fp32 while keeping entire model in fp16.

partially solves #1656
This PR only handles compilation part of multitarget.

Use generate_argument instead of generate_literal for python output as generate_literal doesnt exists
Shorten the names for variables from the main module
Use prefix p_ for parameters
Use shorter variable m for main module in python

* Allow quantizing for both int8 and fp16

Use std::numeric_limits::min/max() functions plus the appropriate value to encode -inf/inf 

Enable retrieving the code object to do tuning in the future.

Refactor supported gfx archs

back out changes for rocm-5.5

* use half hip functions to compute max and min
* add verify test for min and max

Co-authored-by: Paul Fultz II <pfultz2@yahoo.com>

Co-authored-by: Sam Wu <sam.wu2@amd.com>
Co-authored-by: Paul <pfultz2@yahoo.com>

Move CI to support the rocm5.5 release

Adding support for broadcasted scalars to unsqueeze op.

Specifying steps other than 1 is disallowed in this implementation since we want the output the always be a tensor. We can support varying step sizes if we allow a broadcasted scalar output from this op.

Co-authored-by: causten <causten@users.noreply.github.com>

Example of using the C++ API to run an ONNX model with dynamic batch

Remove various file not required for what we use Github runners for

Python API with documentation updates

Adds support for slice,transpose,contigous and reshape fusions into input tensors for a fused mlir kernel.

add option to print tf supported ops
sort both onnx and tf ops alphabetically

When multiplying either the input or output across the K dimensions then the multiple can be applied to the constant which can then be folded with propagate_const.

Exposed the mlir_enabled() call the decide for lowering pipeline's enablement
Disabled the rewrite quantization pipeline in mlir compilation
Added quant convolution as anchor ops
Fixed the return type expectations
Added the fall back hip implementation for quantizelinear and dequantizelinear
Will need advises to improve the implementation for quantizelinear

Relies on Removed split_single_dyn_dim compile flag #1711
Exposes dynamic_dimension as a opaque object with dynamic_dimensions and optimals
Exposes ONNX dyn_input_dims and default_dyn_dim to run with dynamic batch
Updates api.py to be able to create objects from aggregate initialization (used for dynamic_dimension)
Uses offload copy for now

Improves the constant propagation for bert models. Larger batch size no longer use as large of constants.  Also improves the speed of model compilation