[Nuphar] added Gemm-to-MatMul conversion in model editor #4691
Conversation
This file contains 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
* added a mode gemm_to_matmul that turns Gemm Ops into MatMul Ops * enabled model_quantizer to quantize MatMul inside a Loop op * this PR also included Gemm-11 support from Ke Deng
Fixed a couple of old bugs exposed by the newly-added tests and the support of Gemm-11, including: * correctly handle aliasing among states and outputs in Scan * fixed a transpose issue in building tvm IR for MatMul * fixed an issue related to generating IR for computing Gemm alpha * disabled several tests that triggered some deep issue (likely) in the graph partitioner. I think it might be better to have a separate PR to address the issue.
This was referenced May 14, 2022
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Outdated suggestions cannot be applied.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.
added a mode gemm_to_matmul that turns Gemm Ops into MatMul Ops
enabled model_quantizer to quantize MatMul inside a Loop op
this PR also included Gemm-11 support from Ke Deng