Issues

Calibration reader uses the wrong batches after set_range and cannot replay bounded ranges
microsoft/Olive#2688 · opened on Sep 25, 2026
Support fusion-compatible quantization groups for ModuleList MoE experts
microsoft/Olive#2683 · opened on Sep 23, 2026
[Bug]: import olive fails out of the box on 0.13.0 - the telemetry exporter imports requests, which is not a declared dependency
microsoft/Olive#2676 · opened on Sep 20, 2026
[FR]: Docker system - mount input resources read-only, and document/scope the HuggingFace token passed into the container
microsoft/Olive#2675 · opened on Sep 20, 2026
[FR]: Shared-cache integrity hardening (wider model id, source-bound cached configs, normalized blob names) + pin the Miniconda installer fetch
microsoft/Olive#2674 · opened on Sep 20, 2026
Design reproducible coding-agent calibration profiles and manifests
microsoft/Olive#2642 · opened on Aug 27, 2026
[FR]: Enable mobius exporter in optimize cli command
microsoft/Olive#2639 · opened on Aug 26, 2026
Add fused-MoE support to SelectiveMixedPrecision
microsoft/Olive#2638 · opened on Aug 26, 2026
1
Track: revisit/deprecate MobiusBuilder's components_to_export once MobiusBuilder sees wider adoption
microsoft/Olive#2631 · opened on Aug 18, 2026
Design: support quantizing is_transposed=True MoE architectures (gpt-oss, llama4, aria)
microsoft/Olive#2621 · opened on Aug 12, 2026
MoE-quantized models crash on default grouped_mm inference path (QuantTensor transpose not supported)
microsoft/Olive#2619 · opened on Aug 12, 2026
1
Multimodal quantization investigation
microsoft/Olive#2613 · opened on Aug 7, 2026
2
MoE quantization: investigate AWQ (autoawq) support for MoE experts
microsoft/Olive#2600 · opened on Jul 30, 2026
Replace onnxruntime.tools.symbolic_shape_infer with onnx-shape-inference
microsoft/Olive#2560 · opened on Jul 3, 2026
Codebase review: prioritized recommendations across architecture, testing, CI, deps, CLI, docs & performance
microsoft/Olive#2516 · opened on Jun 12, 2026
OnnxKQuantQuantization skips per-expert MatMuls in static-unroll MoE blocks
microsoft/Olive#2489 · opened on Jun 1, 2026
1
Olive feedback: excellent quantization experience with Gemma4 models
microsoft/Olive#2440 · opened on May 4, 2026
olive auto-opt for CPU INT4 fails without --use_model_builder — DynamicCache.from_legacy_cache AttributeError with transformers 5.x
microsoft/Olive#2335 · opened on Feb 16, 2026
3
OOM on 32GB while running auto-opt Qwen/Qwen2.5-Coder-7B-Instruct
microsoft/Olive#2316 · opened on Jan 26, 2026
4
[FR]: Config wizard
microsoft/Olive#2297 · opened on Dec 29, 2025
1
example docs is missing dependencies: ModuleNotFoundError: No module named 'onnxruntime'
microsoft/Olive#2284 · opened on Dec 15, 2025
4
ONNX adapter created by "olive convert-adapters" command cannot work with ONNX model created by "olive auto-opt"
microsoft/Olive#2277 · opened on Dec 5, 2025
Error Unrecognized attribute: qk_output for operator GroupQueryAttention
microsoft/Olive#2276 · opened on Nov 27, 2025
[FR]: Add Support and Documentation for Vision-Language Models (VLMs) like InternVL and Qwen2 VL
microsoft/Olive#2226 · opened on Oct 21, 2025
1
Bug Report: Memory Allocation Error in Qwen2.5-7B/Qwen3-4B Conversion to QNN-Supported ONNX
microsoft/Olive#2210 · opened on Oct 11, 2025
18