Skip to content

Quark GPT-OSS support#1903

Merged
kunal-vaishnavi merged 1 commit into
microsoft:mainfrom
thpereir:main
Dec 29, 2025
Merged

Quark GPT-OSS support#1903
kunal-vaishnavi merged 1 commit into
microsoft:mainfrom
thpereir:main

Conversation

@thpereir
Copy link
Copy Markdown
Contributor

@thpereir thpereir commented Dec 3, 2025

Add Quark GPT-OSS support

  • General support for QMoE zero point/asymmetric quantization
  • New layers used by Quark quantized models
  • Packing used for gate_up proj and down_proj inside Experts

Comment thread src/python/py/models/builders/gptoss.py Fixed
Comment thread src/python/py/models/builders/gptoss.py Fixed
@thpereir thpereir marked this pull request as ready for review December 9, 2025 17:36
@thpereir thpereir force-pushed the main branch 2 times, most recently from 3b7c828 to b1e4a78 Compare December 9, 2025 19:17
@thpereir
Copy link
Copy Markdown
Contributor Author

thpereir commented Dec 9, 2025

@microsoft-github-policy-service agree company="

@microsoft-github-policy-service agree company="AMD"

Comment thread src/python/py/models/builders/base.py
Comment thread src/python/py/models/builders/base.py Outdated
Comment thread src/python/py/models/builders/gptoss.py Outdated
Comment thread src/python/py/models/builders/gptoss.py Outdated
Comment thread src/python/py/models/builders/gptoss.py Outdated
Comment thread src/python/py/models/builders/gptoss.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py
Comment thread src/python/py/models/quantized_model.py
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
Comment thread src/python/py/models/quantized_model.py Outdated
@thpereir
Copy link
Copy Markdown
Contributor Author

@kunal-vaishnavi fixed the two remaining points and rebased again. Please take a look!

@kunal-vaishnavi kunal-vaishnavi enabled auto-merge (squash) December 29, 2025 07:26
- General support for QMoE zero point/asymmetric quantization
- New layers used by Quark quantized models
- Packing used for gate_up proj and down_proj inside Experts modules
auto-merge was automatically disabled December 29, 2025 15:07

Head branch was pushed to by a user without write access

@thpereir
Copy link
Copy Markdown
Contributor Author

Windows arm64 workflow timed out. I force-pushed with a commit message change to re-trigger the workflows (no code change).

@kunal-vaishnavi kunal-vaishnavi enabled auto-merge (squash) December 29, 2025 21:49
@kunal-vaishnavi kunal-vaishnavi merged commit 3be7b36 into microsoft:main Dec 29, 2025
15 checks passed
@thiagocrepaldi
Copy link
Copy Markdown

Great work everybody, and Happy Holidays!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants