feature/onnx-to-tosa #4
Draft
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
convtranspose 1d decomposition
[AutoBump] Merge with 30602ba (Apr 29) (1)
[AutoBump] Merge with fixes of 66e82a9 (Apr 30) (2)
[AutoBump] Merge with 07e645b (May 06) (5)
[AutoBump] Merge with fixes of 85ccac7 [May 2] (4)
[AutoBump] Merge with fixes of e23d8ed (May 12) (6)
[AutoBump] Merge with 328c2f9 (May 13) (7)
[Autobump with fixes] Bump to 5e96d18 (Apr 22)
…ition. The greedy rewriter is generally easier to use, as operations do not need to be marked as illegal Signed-off-by: Jonas Rickert <jonas.rickert@amd.com>
…tern Signed-off-by: Jonas Rickert <jonas.rickert@amd.com>
Signed-off-by: Jonas Rickert <jonas.rickert@amd.com>
…o_tosa Allow lowering of (u)int4 to tosa.
Add also implementation for DepthToSpace DCR mode recomposition
[AutoBump] Merge with 6d728e8 (May 01) (3)
[AutoBump] Merge with 2694354 (May 19) (9)
Bump Quantize/DequantizeLinear to opset 21, adding int4/uint4 support
`getGenOpMix<ONNXRoundOp>` is used in src/Conversion/ONNXToKrnl/Quantization/QuantizeLinear.cpp. There, it can only see the generic definition of the template in `src/Conversion/ONNXToKrnl/ONNXToKrnlCommon.hpp`. But there is also a explicit specialization in src/Conversion/ONNXToKrnl/Math/Elementwise.cpp. To ensure that the explicit specialization is used instead of instantiating the generic template, forward declare the explicit specialization.
…ter) is not constant
…ing_to_tosa_if_size_is_not_constant Prevent ONNX to TOSA legalization when the size is non-constant
…m the model for custom ops. Add a new option runOnnxShapeInference to allow to control it independently of model types. Update default values to be the same in all places.
Add shape inference for microsoft QuickGelu
…m_op Add new option useOnnxModelTypesForCustomOps to allow using types fro…
Fix duplicate symbol getGenOpMix<ONNXRoundOp>
call relevant getter for signed, unsigned and signless ints
Make old opset warning based on newest version in opset range, not oldest
…id_negative_indices [AIESW-13111] Convert negative indices correctly in Gather Simplify Shape
…_for_static_shape [AIESW-13111] Check Gather input shape is static
…anges Allow padValue to be non-constant
…match fixing a mismatch in 4x4 kernel usecase
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
DO NOT merge, just to look at our diff