Narendasan/int8 mixed precision fix #94
Closed
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
overridden Signed-off-by: Naren Dasan <naren@narendasan.com> Signed-off-by: Naren Dasan <narens@nvidia.com>
Adds evaluators for: - aten::eq - aten::ne - aten::lt - aten::gt - aten::le - aten::ge - aten::add - aten::sub - aten::mul - aten::Bool - aten::Float - aten::__not__ - aten::__is__ - aten::__isnot__ - aten::numel - aten::dim - aten::div - aten::floordiv - aten::floor - aten::warn - prim::min - prim::max - prim::shape - prim::unchecked_cast - prim::Uninitalized - prim::RaiseException Signed-off-by: Naren Dasan <naren@narendasan.com> Signed-off-by: Naren Dasan <narens@nvidia.com>
Signed-off-by: Naren Dasan <naren@narendasan.com> Signed-off-by: Naren Dasan <narens@nvidia.com>
Signed-off-by: Naren Dasan <naren@narendasan.com> Signed-off-by: Naren Dasan <narens@nvidia.com>
int8 do not enable fp16 kernels Signed-off-by: Naren Dasan <naren@narendasan.com> Signed-off-by: Naren Dasan <narens@nvidia.com>
github-actions
bot
added
component: conversion
mfeliz-cruise
added a commit
to mfeliz-cruise/Torch-TensorRT
that referenced
this pull request
Jan 5, 2023
…g to float (pytorch#94) * Support int inputs to aten::max/min and aten::argmax/argmin by casting to float * correct layer name * address nit, remove local variable
mfeliz-cruise
added a commit
to mfeliz-cruise/Torch-TensorRT
that referenced
this pull request
Jan 9, 2023
…g to float (pytorch#94) * Support int inputs to aten::max/min and aten::argmax/argmin by casting to float * correct layer name * address nit, remove local variable
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Outdated suggestions cannot be applied.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.
Description
When int8 is enabled, we also need to enable fp16 kernels just in case those kernels provide a faster path. This explains significant performance differences between FP16 and INT8 in cases where most layers are not mapped to INT8 kernels.
Partially addresses #93
Type of change
Please delete options that are not relevant and/or add your own.
Checklist:
make htmlin docsrc)