microsoft/Olive

View on GitHub

Olive: Simplify ML Model Finetuning, Conversion, Quantization, and Optimization for CPUs, GPUs and NPUs.

Stars: 2.4k
Forks: 315
Open beginner issues: 0
Indexed issues: 29
Avg merge: 4d 11h
Merged PRs (30d): 16

Dominant language: Python
License: MIT
Last GitHub push: Sep 15, 2026
Latest indexed: Sep 19, 2026
Contributing guide: Contributing guide
Code of conduct: Code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

29 open issues indexed

Design reproducible coding-agent calibration profiles and manifests Open

microsoft/Olive#2642 · 1 assignee · Opened Aug 27, 2026
[FR]: Enable mobius exporter in `optimize` cli command Open

enhancement

microsoft/Olive#2639 · 1 reaction · 2 assignees · Opened Aug 26, 2026
Add fused-MoE support to SelectiveMixedPrecision Open

microsoft/Olive#2638 · 1 comment · 1 assignee · Opened Aug 26, 2026
Track: revisit/deprecate MobiusBuilder's components_to_export once MobiusBuilder sees wider adoption Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

microsoft/Olive#2631 · Opened Aug 18, 2026
Design: support quantizing is_transposed=True MoE architectures (gpt-oss, llama4, aria) Open

microsoft/Olive#2621 · 1 assignee · Opened Aug 12, 2026
MoE-quantized models crash on default `grouped_mm` inference path (QuantTensor transpose not supported) Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 42/100

microsoft/Olive#2619 · 1 comment · Opened Aug 12, 2026
Multimodal quantization investigation Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

microsoft/Olive#2613 · 2 comments · Opened Aug 7, 2026
MoE quantization: investigate AWQ (autoawq) support for MoE experts Open

microsoft/Olive#2600 · 1 assignee · Opened Jul 30, 2026
Replace onnxruntime.tools.symbolic_shape_infer with onnx-shape-inference Open

microsoft/Olive#2560 · 1 reaction · 2 assignees · Opened Jul 3, 2026
Codebase review: prioritized recommendations across architecture, testing, CI, deps, CLI, docs & performance Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

microsoft/Olive#2516 · 1 comment · Opened Jun 12, 2026
OnnxKQuantQuantization skips per-expert MatMuls in static-unroll MoE blocks Open

Difficulty 5/5 Over a week Newbie friendliness 35/100

microsoft/Olive#2489 · 1 comment · Opened Jun 1, 2026
Olive feedback: excellent quantization experience with Gemma4 models Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

microsoft/Olive#2440 · Opened May 4, 2026
`olive auto-opt` for CPU INT4 fails without `--use_model_builder` — `DynamicCache.from_legacy_cache` AttributeError with transformers 5.x Open

Difficulty 4/5 3-5 days Newbie friendliness 45/100

microsoft/Olive#2335 · 3 comments · Opened Feb 16, 2026
OOM on 32GB while running auto-opt Qwen/Qwen2.5-Coder-7B-Instruct Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

microsoft/Olive#2316 · 4 comments · Opened Jan 26, 2026
[FR]: Config wizard Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 35/100

microsoft/Olive#2297 · 1 comment · Opened Dec 29, 2025
example docs is missing dependencies: ModuleNotFoundError: No module named 'onnxruntime' Open

Difficulty 2/5 1-3 hours Newbie friendliness 35/100

microsoft/Olive#2284 · 4 comments · Opened Dec 15, 2025
ONNX adapter created by "olive convert-adapters" command cannot work with ONNX model created by "olive auto-opt" Open

Difficulty 4/5 3-5 days Newbie friendliness 42/100

microsoft/Olive#2277 · Opened Dec 5, 2025
Error Unrecognized attribute: qk_output for operator GroupQueryAttention Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

microsoft/Olive#2276 · Opened Nov 27, 2025
[FR]: Add Support and Documentation for Vision-Language Models (VLMs) like InternVL and Qwen2 VL Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 20/100

microsoft/Olive#2226 · 1 comment · Opened Oct 21, 2025
Bug Report: Memory Allocation Error in Qwen2.5-7B/Qwen3-4B Conversion to QNN-Supported ONNX Open

Difficulty 4/5 3-5 days Newbie friendliness 30/100

microsoft/Olive#2210 · 18 comments · Opened Oct 11, 2025
Sample Question - Error when Converting Phi 3.5 to QNN Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

microsoft/Olive#2200 · 1 comment · Opened Oct 10, 2025
[FR]: Simplified Model Architecture Integration via Class-Based API Open

enhancement

microsoft/Olive#2130 · 1 assignee · Opened Sep 1, 2025
Add support for zai-org/GLM-4.1V-9B-Thinking Open

model support request

Difficulty 4/5 3-5 days Newbie friendliness 35/100

microsoft/Olive#2118 · Opened Aug 21, 2025
According to the example content, the model has been successfully exported, but Cannot run in Net environment Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

microsoft/Olive#2103 · 4 comments · Opened Aug 16, 2025
Phi-3.5-mini QNN example broken by latest transformers Open

microsoft/Olive#2051 · 1 comment · 1 assignee · Opened Aug 4, 2025
Executing Olive for QNN EP for llama3.2 1B instruct, getting Invalid engine (type=value_error) Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

microsoft/Olive#1905 · 7 comments · Opened Jun 10, 2025
Will QNN-EP optimization work for a non X series snapdragon device? Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

microsoft/Olive#1890 · 4 comments · Opened Jun 2, 2025
[FR]: Export graphs in NHWC format Open

enhancement

Difficulty 5/5 Over a week Newbie friendliness 25/100

microsoft/Olive#1791 · Opened Apr 23, 2025
[FR]: Conda recipe Open

enhancement

Difficulty 3/5 1-2 days Newbie friendliness 48/100

microsoft/Olive#1537 · Opened Jan 6, 2025

microsoft/Olive

Get new issues in your inbox