Skip to content

Pull requests: microsoft/onnxruntime-genai

New pull request New
90 Open 1,814 Closed
90 Open 1,814 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Infer multimodal feature width from model I/O shape
#2608 opened Sep 22, 2026 by Xiaoyu Z (xiaoyu-work) Member Loading…
1
fix: size fp32 logits buffer with last-token shape during prefill
#2607 opened Sep 22, 2026 by Zhong, GenMing (amd-genmingz) Loading…
2
test qwen 3.8 webgpu
#2606 opened Sep 21, 2026 by aciddelgado Contributor • Draft
fix: allocate multimodal pipeline buffers on the session that writes them
#2605 opened Sep 21, 2026 by Yuri Khrustalev (ykhrustalev) Contributor Loading…
1
8
Ubuntu 24.04 image override
#2602 opened Sep 21, 2026 by kpkbandi • Draft
feat: add LFM2-Audio / LFM2.5-Audio speech input and output support
#2601 opened Sep 20, 2026 by Yuri Khrustalev (ykhrustalev) Contributor Loading…
9
feat(builder): support INT2 DFlash 2 drafters
#2600 opened Sep 20, 2026 by Tianlei Wu (tianleiwu) Contributor Loading…
Standardize telemetry events and dependency policy
#2599 opened Sep 20, 2026 by bmehta001 Contributor Loading…
@bmehta001
3
Reduce intermediate memory in quantized weight packing
#2596 opened Sep 20, 2026 by 野生の男 (Yasei-no-otoko) Loading…
2
Load model builders and Transformers architectures lazily
#2594 opened Sep 18, 2026 by xieofxie Loading…
5
Feat: Add safetensors LoRA adapter loading to the C model benchmark
#2593 opened Sep 18, 2026 by Chang Liu (cliu1003) Contributor Loading…
4 tasks
6
Add structured model builder configuration
#2592 opened Sep 18, 2026 by Tianlei Wu (tianleiwu) Contributor Loading…
4 tasks done
71
Adding word level timestamps to Nemotron Streaming Pipeline
#2591 opened Sep 17, 2026 by Mason Corey (themason2011) Loading…
8
Centralize packed position-ID plane detection and validate generated ranges
#2589 opened Sep 17, 2026 by aciddelgado Contributor Loading…
1
[WIP] Fix CUDA illegal memory access in PagedAttention on full-context requests
#2587 opened Sep 17, 2026 by Copilot AI • Draft
1 of 4 tasks
1
@tianleiwu
1
Rebuild AMDGPU device allocator per model instead of caching across models
#2584 opened Sep 17, 2026 by KenLagos Loading…
3
Add persistent parallel preprocessing loops
#2582 opened Sep 16, 2026 by Copilot AI Loading…
@kunal-vaishnavi
14
Update Windows GPU pool to new scaleset.
#2577 opened Sep 16, 2026 by eserscor Contributor • Draft
Try migrate to scaleset
#2574 opened Sep 16, 2026 by eserscor Contributor Loading…
1
Add prefix caching to the dynamic engine
#2573 opened Sep 16, 2026 by bmehta001 Contributor Loading…
@bmehta001
40
[WebGPU] Support paged attention models
#2570 opened Sep 15, 2026 by Prathik Rao (prathikr) Contributor Loading…
12
Add an API to release cached device resources after model unload
#2569 opened Sep 15, 2026 by aciddelgado Contributor Loading…
4
Add a native Gemma 3n multimodal processor
#2567 opened Sep 15, 2026 by Anirudh Swaminathan (Anirudh-Swaminathan) Loading…
16
[Draft] GPT-OSS: opt-in fused com.migraphx GptOssMoE INT4 MoE for the MIGraphX plugin-EP path
#2562 opened Sep 14, 2026 by rlegithub • Draft
2
Add Muse Glimmer (Llama-4-family) OGA model builder + muse_glimmer model type
#2561 opened Sep 14, 2026 by rlegithub • Draft
1 task
4
ProTip! Follow long discussions with comments:>50.