Skip to content
This repository has been archived by the owner on Oct 25, 2024. It is now read-only.

Add weight_only support for PyTorch framework #297

Merged
merged 25 commits into from
Sep 15, 2023
Merged

Add weight_only support for PyTorch framework #297

merged 25 commits into from
Sep 15, 2023

Conversation

PenghuiCheng
Copy link
Contributor

Type of Change

feature
No API changed

Description

Add weight_only support for PyTorch framework

Expected Behavior & Potential Risk

the expected behavior that triggered by this PR

How has this PR been tested?

Quantized model with weight-only quantization successfully.

PenghuiCheng and others added 20 commits September 5, 2023 09:24
@PenghuiCheng
Add weight_only support for PyTorch framework
2e35c81 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@zhewang1-intc
update isa check on no-amx-supported platform
268e068
@PenghuiCheng
Fixed building issue and get_version issue
a8c09bc 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Fixed UT error
1246d55 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Add weight_only support for PyTorch framework
2b29630 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@zhewang1-intc @PenghuiCheng
update isa check on no-amx-supported platform
8df521f
@PenghuiCheng
Fixed building issue and get_version issue
39096e9 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Fixed UT error
f69d252 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@VincyZhang
fix pylint
25f68b9 
Signed-off-by: Wenxin Zhang <wenxin.zhang@intel.com>
@VincyZhang
Merge branch 'penghuic/qbits_porting' of https://github.com/intel/int…
7495c1f 
…el-extension-for-transformers into penghuic/qbits_porting
@PenghuiCheng
Fixed pylint error
26b90e7 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
merge main branch
ef70865 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Fixed pylint error
75de4d6 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@VincyZhang
Update install_binary.sh
f3611d0
@PenghuiCheng
Merge main branch
2819294 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Merge main branch
9a244aa 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Fixed UT error with master version neural-compressor
7b2834a 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Merge remote-tracking branch 'origin' into penghuic/qbits_porting
74e76b9
@PenghuiCheng
Add UT for weight-only quantization
4dd5d55 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Remove modeling_causal.py since modeling_auto replace it
bec8576 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
zhewang1-intc
zhewang1-intc reviewed Sep 12, 2023
View reviewed changes
Copy link
Contributor

@zhewang1-intc zhewang1-intc left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

consider to sync with the latest QBits backend?

@VincyZhang
Copy link
Contributor

extension test：https://inteltf-jenk.sh.intel.com/view/nlp-toolkit-validation/job/nlp-toolkit-validation-top-mr-extension/1310/

@hshen14
Copy link
Contributor

hshen14 commented Sep 13, 2023

Add a dependency: #297

@VincyZhang VincyZhang added the WIP label Sep 14, 2023
DDEle
DDEle reviewed Sep 14, 2023
View reviewed changes
intel_extension_for_transformers/llm/runtime/deprecated/third_party/oneDNN
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Unexpected change?

@PenghuiCheng
Fixed import error when loading libweight_only_jblasop.so
77af939 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
merge main branch
79a1842 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Update jblass and weight-only UT
dd3ebfb 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@PenghuiCheng
Support gcc 8.4 version to build jblass
96180e0 
Signed-off-by: Cheng, Penghui <penghui.cheng@intel.com>
@VincyZhang
Copy link
Contributor

Coverage Drop
Unit Test failed with lines coverage decrease -2.310%
Unit Test failed with branches coverage decrease -4.126%
please add UT to cover newly-added modules. @PenghuiCheng

@VincyZhang VincyZhang added ITREX v1.2 and removed WIP labels Sep 15, 2023
@VincyZhang VincyZhang merged commit 3a064fa into main Sep 15, 2023
@VincyZhang VincyZhang deleted the penghuic/qbits_porting branch September 15, 2023 10:58
@changwangss changwangss mentioned this pull request Sep 15, 2023
Merged
@zhenwei-intel zhenwei-intel mentioned this pull request Sep 18, 2023
Merged
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Reviewers

@zhewang1-intc zhewang1-intc zhewang1-intc left review comments

@DDEle DDEle DDEle left review comments

@VincyZhang VincyZhang VincyZhang approved these changes

@xin3he xin3he xin3he approved these changes

@changwangss changwangss Awaiting requested review from changwangss

@kevinintel kevinintel Awaiting requested review from kevinintel

@a32543254 a32543254 Awaiting requested review from a32543254

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
6 participants
@PenghuiCheng @VincyZhang @hshen14 @DDEle @zhewang1-intc @xin3he