onnxruntime
[Migraphx EP] Static int8 QDQ support
#17931
Merged
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Overview
Commits
33
Changes
View On
GitHub
[Migraphx EP] Static int8 QDQ support
#17931
PeixuanZuo
merged 33 commits into
microsoft:main
from
TedThemistokleous:migraphx_int8_support
TedThemistokleous
marked this pull request as ready for review
2 years ago
github-advanced-security
commented on 2023-10-18
github-advanced-security
commented on 2023-10-19
github-advanced-security
commented on 2023-10-20
github-advanced-security
commented on 2023-10-20
TedThemistokleous
force pushed
from
35225d12
to
b082a99e
2 years ago
TedThemistokleous
changed the title
Migraphx Static int8 QDQ support
[Migraphx EP] Static int8 QDQ support
2 years ago
cloudhan
requested a review
from
justinchuby
2 years ago
cloudhan
requested a review
from
cloudhan
2 years ago
cloudhan
requested a review
from
PeixuanZuo
2 years ago
Add flags for quantize int8
f196e5a1
Add pieces to grab calibration data for int8
6ca1f189
Remove calibration table things for now
1dd107cd
Add in quantization step
5853dcb2
update benchmark.py in transformers to allow int8 runs
e5f128df
Fix naming for environment env kINT8Enable
8a9555a3
Backup changes - int8 + static calibration table hooks
269f5bcb
Update pybind api to allow for MIGraphX EP options to be set
26461167
Add Verbose options for MIGraphX on EP creation
e84cb585
Use ReadDynamicRange from TensorRT EP to reuse cal table for static i…
1d1fb222
Include flatbuffer file generated for TRT calibration data types
cf31985f
Add hooks for calibration table similar to TensorRT EP
f08e930e
Add reading in of calibration table and push to migraphx calibration
299279a8
Add support for QLinearAdd, QLinearConv, QLinearMatMul
6d1bfa67
Fix initial calibration based on compile
c6f47e1f
Lint/Formatting
8f4d63ab
Add missing changes for provider info
978c8149
gate benchmark.py to migraphx ep for int8
610a6a67
Remove MIGraphx from training module
3f68a49c
Fix provider define for python lint
7db7fd83
Fix format
5a7954b5
Remove onnxruntime.quantization from benchmark.py
1675d4f4
Fixing linting/formatting and more issues.
9e449abc
fixup! Fixing linting/formatting and more issues.
cd8d4339
Lintrunner pass
63cce29c
Set convert to uint32_t instead of uint64_t
5feabdfb
Update things that are breaking builds for MIGraphX EP
b6df9456
Further Fixes for C++ lint
82c4178d
Additional lintrunner pass - disregarding C++ linter pass
219b513a
Fix typo.
ccc4fdbe
Fix CUDA comment
938dd455
Final lintrunner pass for CI
7c51b947
TedThemistokleous
force pushed
from
9e67abaf
to
7c51b947
2 years ago
jeffdaily-ms
dismissed these changes on 2023-11-08
Fix additional lint errors.
037d9285
TedThemistokleous
dismissed their stale review via
037d9285
2 years ago
TedThemistokleous
requested a review
from
jeffdaily-ms
2 years ago
PeixuanZuo
approved these changes on 2023-11-09
PeixuanZuo
merged
8d503138
into main
2 years ago
TedThemistokleous
deleted the migraphx_int8_support branch
2 years ago
Login to write a write a comment.
Login via GitHub
Reviewers
PeixuanZuo
jeffdaily-ms
github-advanced-security
justinchuby
cloudhan
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub