-
Notifications
You must be signed in to change notification settings - Fork 1.6k
Pull requests: jundot/omlx
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(scheduler): drop the prefix-cache hit after a sparse-prefill failure
#2484
opened Aug 2, 2026 by
JimStenstrom
Contributor
Loading…
Complete Simplified Chinese localization for the native app and web console
#2481
opened Aug 2, 2026 by
LXD-8
Loading…
feat：MTPLX / mlx-lm-forge Model Exported MTP (Multi-Head Prediction) Nativ…
#2478
opened Aug 2, 2026 by
HelloTianxu
Loading…
DeepSeek-V4-Flash-0731: embedded DSpark speculative decoding on the stock HF checkpoint (env-gated, c=1)
#2468
opened Aug 2, 2026 by
yohann-bearzi
Contributor
Loading…
chore(deps): bump markitdown from 0.1.6 to 0.1.7
dependencies
Pull requests that update a dependency file
python
Pull requests that update python code
#2466
opened Aug 2, 2026 by
dependabot
Bot
Loading…
fix: do not recover truncated thinking as content on the streaming paths
#2458
opened Aug 1, 2026 by
mabaeyens
Loading…
feat(vlm_mtp): enforce thinking budget inside MTP speculative decoding
#2456
opened Aug 1, 2026 by
Maxim-Mazurok
Loading…
fix(reranker): support Jina v3.5 reranker projector key naming
#2449
opened Aug 1, 2026 by
damascoemi
Loading…
feat: add scheduler-native dSpark speculative decoding
#2441
opened Jul 31, 2026 by
wd041216-bit
•
Draft
fix(server): null-guard the specprefill model-settings fallback
#2437
opened Jul 31, 2026 by
JimStenstrom
Contributor
Loading…
fix(oq): clamp the sensitivity perturbation to a valid affine group size
#2433
opened Jul 30, 2026 by
JimStenstrom
Contributor
Loading…
fix(discovery): skip incompletely downloaded sharded models
#2432
opened Jul 30, 2026 by
JimStenstrom
Contributor
Loading…
fix(engine_pool): abort in-flight requests before manual engine unload
#2429
opened Jul 30, 2026 by
JimStenstrom
Contributor
Loading…
fix: credit prefill preflight cached_tokens only for hot-cache-verified blocks
#2427
opened Jul 30, 2026 by
fparrav
Contributor
Loading…
3 tasks done
fix(api): preserve partial prefill flag through message drop/merge (#2357)
#2426
opened Jul 30, 2026 by
richgoodson
Contributor
Loading…
Distributed cluster serving across Macs (tensor + pipeline parallel)
#2423
opened Jul 30, 2026 by
ashhart
Loading…
i18n(app): Add zh-Hans & zh-Hant translations to macOS app + fix Swift 6.1 build errors
#2419
opened Jul 30, 2026 by
NirvanaCh7
Contributor
Loading…
feat: add durable idempotency lookup for Responses API
#2397
opened Jul 28, 2026 by
andyelka-creator
Loading…
perf(cache): store large fp32 boundary-snapshot states as int8
#2366
opened Jul 25, 2026 by
davidtorcivia
Contributor
Loading…
feat(nemotron_h): fused Metal decode-step kernels + per-position SDPA for verify windows
#2365
opened Jul 25, 2026 by
davidtorcivia
Contributor
Loading…
feat(specprefill): MLA query extractor for DeepSeek/GLM-family drafts
#2355
opened Jul 23, 2026 by
StartupBros
Contributor
Loading…
Previous Next
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.