Skip to content
Navigation Menu
Sign in
Products
Platform
GitHub
Repository hosting, code review, and collaboration
Copilot
AI coding agent for tasks, projects, and automated work
Build & Ship
Actions
CI/CD and automation
Code Quality
Quality analysis, fixes, and enforcement
Codespaces
Cloud development environments
Secure
Code Security
Vulnerability detection and remediation
Secret Protection
Secret scanning and leak prevention
All Features
Developers
Learn
Documentation
Training
Certifications
Build
Build with Copilot
Customize Copilot
Integrations
Explore
Trending repos
GitHub Sponsors
Developer articles
Changelog
Changelog
Your source for the latest features, improvements, and fixes across GitHub products, all in one place.
Businesses
By Stage
Enterprises
Team
Startups
Nonprofits
By Industry
Healthcare
Financial services
Manufacturing
Government
See all industries
Evaluate GitHub
Why GitHub
Customer Stories
Business insights
Trust Center
Add-ons
GitHub Advanced Security
Enterprise-grade code security
Copilot for business
Enterprise-grade AI features
Premium Support
Enterprise-grade 24/7 support
Resources
Learn & Connect
Documentation
Resource center
Blog
Events & webinars
Community Forum
Customer Stories
Integrations
Support
Customer support
Training
GitHub Status
Company
About
Partners
Changelog
GitHub Universe
GitHub Universe
The event for developers, by developers — talks, workshops, and what’s next for GitHub.
Pricing
Search
/
Sign in
Sign up
Appearance settings
You signed in with another tab or window.
Reload
to refresh your session.
You signed out in another tab or window.
Reload
to refresh your session.
You switched accounts on another tab or window.
Reload
to refresh your session.
Dismiss alert
{{ message }}
vllm-project
vllm-torchtpu
Repository navigation
Code
Issues
125
(125)
Pull requests
892
(892)
Actions
Projects
Security and quality
Insights
More
items
All pull requests
Labels
18
(18)
Milestones
0
(0)
New pull request
Search pull requests
is
:
pr
state
:
open
is:pr state:open
Clear filter
Search
Pull requests
Open
892
(892)
Closed
2,176
(2,176)
Author
Label
Projects
Milestones
Reviews
Assignee
Sort by
Newest
descending
Sort…
Sort by Newest descending
More items
Comfortable display density
Compact display density
[Core][TorchTPU API] Clean up deprecated torch.tpu.current_device mocks in model tests
#3340
·
yilingyuan
opened
Nov 16, 2026
Collaborator
·
5 of 13 tasks
·
·
[CI/Build] Add TPU v6e support across tests, integration, and perf lanes
ready
#3339
·
charlesli640
opened
Nov 15, 2026
Collaborator
·
1 of 8 tasks
·
·
2
[glm-agentx-pd][Misc] Worker: log host memory from one worker per host (TPU_HOST_MEMORY_LOG_S)
#3338
·
PrimaldevX
opened
Nov 15, 2026
·
3 of 13 tasks
·
·
[glm-agentx-pd][Kernel] MLA KV writers: find each row's sequence with a compare, not a while loop
#3337
·
PrimaldevX
opened
Nov 15, 2026
·
3 of 13 tasks
·
·
[glm-agentx-pd][Kernel] Sparse MLA: allow mixed live widths in multiples of the 64-row gather atom
#3336
·
PrimaldevX
opened
Nov 15, 2026
·
3 of 13 tasks
·
·
[glm-agentx-pd][Kernel] moe_tp_decode: serve packed FP4 routed experts in decode and sequence-parallel EP prefill
#3335
·
PrimaldevX
opened
Nov 15, 2026
·
3 of 13 tasks
·
·
[glm-agentx-pd][Kernel][Quantization] Requantize FP8 routed experts to FP4 at load
#3334
·
PrimaldevX
opened
Nov 15, 2026
·
3 of 13 tasks
·
·
[Model][GLM-5.3] Add Glm53DecodeOwner megakernel adapter plugin and shorthand registration
#3333
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
7 of 13 tasks
·
·
[Kernel][GLM-5.3] Add fused 46-layer decode + EAGLE3 MTP Pallas megakernel and weight loader
#3332
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Core] Add model-agnostic DecodeOwner contract, DecodeOwnerScheduler, and runner hooks
#3331
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Model][GLM-5.3] Add Glm5NextMTP EAGLE3 speculative decoding draft model
#3330
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
7 of 13 tasks
·
·
[Model][GLM-5.3] Add GLM-5.3-Flash (Glm5NextForCausalLM, Glm5NextForConditionalGeneration)
#3329
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Kernel][GLM-5.3] Add Glm53TpuKimiDeltaAttention unified-pool KDA layer
#3328
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Kernel][GLM-5.3] Add one-dispatch FP8 MoE custom op and sequence-parallel TP collectives
#3327
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Kernel][GLM-5.3] Add Pallas manifold-constrained hyper-connection (mHC) kernels and custom ops
#3326
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Kernel] Bound pooled KDA state gather/scatter to live rows with explicit row packing
#3325
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Kernel] Support dynamic rhs_group_base indexing for stacked expert stores in gmm_v2
#3324
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[Kernel] Fit Ragged Paged Attention v3 default block sizes within VMEM for wide heads
#3323
·
gitrajav
opened
Nov 15, 2026
Collaborator
·
6 of 13 tasks
·
·
[MoE Router Replay] Add TPU routed-experts capturer
ready
#3322
·
haotianxue-google
opened
Nov 15, 2026
Collaborator
·
·
4
[Structured Decoding][SpecDecode] Implement Grammar-as-a-Draft (GaaD) on v0.33.0
ready
#3321
·
maxwillzq
opened
Nov 15, 2026
Collaborator
·
·
[Kernel][Model] Subspace-RPA for Gemma 4 global attention on TPU
#3320
·
prishajain1
opened
Nov 15, 2026
Collaborator
·
·
[Kernel][Model] Hierarchical Pallas collectives and fused sublayer bridge for Gemma 4
#3319
·
prishajain1
opened
Nov 15, 2026
Collaborator
·
·
[Omni][TPU] Offload QwenImage text_encoder to CPU outside prompt encoding
#3317
·
Perseus14
opened
Nov 14, 2026
Collaborator
·
·
10
[Omni][Kernel] Add Pallas Split-Splash attention in TpuSDPAImpl
#3316
·
Perseus14
opened
Nov 14, 2026
Collaborator
·
·
16
[DSv4.1 Flash] Run short MoE tiles in the 1,024-token build and build the MoE routing plan in one kernel #xClimb
#3315
·
bigeye
opened
Nov 14, 2026
Collaborator
·
·
Previous
1
2
3
4
5
6
7
…
36
Next
You can’t perform that action at this time.