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 }}
mw00
project-maya
Repository navigation
Code
Issues
106
(106)
Pull requests
124
(124)
Discussions
Actions
Security and quality
Insights
More
items
All pull requests
Labels
20
(20)
Milestones
0
(0)
New pull request
Search pull requests
is
:
pr
state
:
open
is:pr state:open
Clear filter
Search
Pull requests
Open
124
(124)
Closed
520
(520)
Author
Label
Projects
Milestones
Reviews
Assignee
Sort by
Newest
descending
Sort…
Sort by Newest descending
More items
Comfortable display density
Compact display density
GLM: an expert pool that cannot be complete takes what its budget holds (MAYA_GLM_POOL_FILL=1, opt-in)
#1148
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
GLM decode: the shared expert and the KDA low-rank projections on a side stream (MAYA_GLM_SHEXP_SIDE=1, opt-in)
#1147
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
GLM: a logit dump in the fast path, for comparing two builds bit for bit (MAYA_GLM_DUMP_LOGITS)
#1146
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
Sampler: the sampled chain's per-entry work in parallel, its vocabulary passes over several blocks
#1145
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
maya-ik2: IK2_BATCH_PREFILL - a prompt chunk's routed experts batched (from alesha-pro/maya-inkling-mini)
#1144
·
mrexodia
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
Loop guard: keep what it works out of each message's own text (MAYA_LOOP_GUARD_MEMO, opt-in)
#1142
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
DeepSeek: the GEN line carries only a prompt's new ids (MAYA_GEN_KEEP + MAYA_DS4_GEN_KEEP, opt-in)
#1141
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
DeepSeek decode: a window's router gate GEMV on the one-token kernel (DS4_ROUTER_ROWS=2, opt-in)
#1140
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
GLM decode: a window's tokens take the MAYA_GLM_DSA_SCORE2 kernel too
#1139
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
GLM decode: the MLA split attention reads INT8 latent records 16 values a load
#1138
·
vdkcf
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
GLM DFlash decode: the first answer token after a forced reasoning close continued the reasoning
#1137
·
vdkcf
opened
Nov 22, 2026
·
3 of 5 tasks
·
·
maya-ik2 on AMD under Windows: built and run on a Radeon AI PRO R9700
#1135
·
HermesVC
opened
Nov 22, 2026
·
·
serve: Inkling 2's tool calls stream while the model writes them
#1134
·
HermesVC
opened
Nov 22, 2026
·
·
maya-ik2: 8 more expert block types, CPU expert lane, offload deadlock fix
#1133
·
TokenTinkererAI
opened
Nov 22, 2026
·
·
mm3(hip/gfx906): add MAYA_MM3_DSPARK_FC16=1 FP16 draft-fc fallback for MiniMax-M3.1
#1132
·
nicholas631
opened
Nov 22, 2026
Contributor
·
5 of 6 tasks
·
·
maya-ik2: a session cache in host RAM - conversations taking turns resume without re-reading their prompts
#1130
·
mrexodia
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
serve: Inkling 2's "minimal" effort reaches its template as 0.1, not off
#1129
·
mrexodia
opened
Nov 22, 2026
·
4 of 5 tasks
·
·
Windows: find Visual Studio with a standalone vswhere on PATH
#1128
·
Thymos2027
opened
Nov 22, 2026
·
6 tasks done
·
·
Hy4-Flash in the web app: the Saved line, VRAM hit, what fills the GPUs, and where the experts live from the start
#1127
·
needmorevram
opened
Nov 22, 2026
Contributor
·
·
Keep manual GLM Flash tuning when rerunning CUDA setup
#1125
·
oscarlius
opened
Nov 22, 2026
Contributor
·
·
maya-hy4: restore points (--snap-every N, off by default)
#1123
·
lij768423-svg
opened
Nov 22, 2026
Contributor
·
4 of 5 tasks
·
·
DeepSeek: show prefill VRAM hit rate separately from decode
#1119
·
LostVector
opened
Nov 22, 2026
Contributor
·
·
ik2: INT4 Q4_32 KV + host streaming & GPU dynamic Clock-LFU offload cache (41+ tok/s @ 128K on 8 GB RTX 3070)
#1118
·
saatvik-viniak
opened
Nov 22, 2026
Contributor
·
·
DeepSeek: reuse busy conversation state for parallel prompt continuations
#1117
·
LostVector
opened
Nov 22, 2026
Contributor
·
·
tools/ik2: a modified checkpoint (a derisked finetune) as an inkling2 GGUF for maya-ik2
#1115
·
CDC6600
opened
Nov 22, 2026
·
·
Previous
1
2
3
4
5
Next
You can’t perform that action at this time.