-
Notifications
You must be signed in to change notification settings - Fork 1.2k
Pull requests: jundot/omlx
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Fix admin chat preserve_thinking history serialization
#1329
opened May 20, 2026 by
gindinconsulting
Loading…
fix(tests): prevent engine loop from calling step() in abort/cancelation tests
#1324
opened May 20, 2026 by
fry69
Contributor
Loading…
fix(load): wire MTP sanitize-preservation patch into VLM inference path
#1320
opened May 20, 2026 by
sangemaru
Loading…
fix(engine): per-engine threads to eliminate cross-engine stream contamination
#1304
opened May 19, 2026 by
ivaniguarans
Contributor
Loading…
Fix #1249: Handle unexpected 'strict' kwarg in mlx_lm load()
#1297
opened May 18, 2026 by
zyguy
Loading…
Eval Reliability, Benchmark History, Model Manager QoL, and MTP Batch Safety
#1254
opened May 14, 2026 by
aljen
Loading…
feat(oq): extend calibration data with code, tool-calling, and reasoning
#1246
opened May 13, 2026 by
a4501150
Contributor
Loading…
4 tasks done
feat(chat): add temperature control, persist and ui
#1213
opened May 13, 2026 by
beamivalice
Contributor
Loading…
fix(admin/benchmark): forward SpecPrefill model settings to engine
#1191
opened May 11, 2026 by
SuperMarioYL
Loading…
2 of 3 tasks
feat(bench): inline run-time settings modal on bench tabs
#1187
opened May 11, 2026 by
jasonpaulso
Contributor
Loading…
9 tasks done
feat(dflash): Path A double-engine + Gemma 4 backend — generalized for Qwen + Gemma 4
#1185
opened May 11, 2026 by
panwudi
Loading…
10 of 12 tasks
2
feat(tool_calling): parse Llama-3-style {"name","parameters"} JSON content
#1153
opened May 9, 2026 by
aeyeopsdev
Loading…
fix: catch TypeError when accessing think token properties
#1141
opened May 9, 2026 by
nightguarder
Loading…
docs: add xgrammar structured output troubleshooting section and fix script
#1075
opened May 5, 2026 by
scubamount
Loading…
fix(dflash): route image-bearing requests to the VLM fallback engine
#1056
opened May 4, 2026 by
0xAlcibiades
Loading…
feat(bench): Add opt-out toggle for community benchmark upload
#1052
opened May 4, 2026 by
jabagawee
Loading…
fix: treat lfm2 as ambiguous model type (LLM + embedding + audio)
#1036
opened May 1, 2026 by
unstoppablesssss
Loading…
Add Darwin Nix flake to Allow import within nix-darwin
#1026
opened May 1, 2026 by
dzmitry-lahoda
Loading…
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.