-
Notifications
You must be signed in to change notification settings - Fork 111
Pull requests: lightseekorg/smg
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(grpc): resolve string stop sequences for SGLang skip_tokenizer_init workers
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
#1877
opened Jul 4, 2026 by
gongwei-130
Collaborator
Loading…
fix(grpc/chat): honor reasoning_effort "none"/"minimal" as enable_thinking=false
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
priority:high
High priority
#1876
opened Jul 3, 2026 by
qywu
Contributor
Loading…
perf(multimodal): optimize Qwen vision preprocessing
dependencies
Dependency updates
multimodal
Multimodal crate changes
priority:high
High priority
#1875
opened Jul 3, 2026 by
yechank-nvidia
Collaborator
Loading…
2 of 4 tasks
perf(multimdoal): Optimize OpenCV video decoding and thread allocation
dependencies
Dependency updates
multimodal
Multimodal crate changes
priority:high
High priority
#1865
opened Jul 2, 2026 by
yechank-nvidia
Collaborator
Loading…
fix(cache-aware): use backend global backlog for load-aware fallback
model-gateway
Model gateway crate changes
needs-rebase
PR has merge conflicts that need to be resolved
protocols
Protocols crate changes
#1863
opened Jul 1, 2026 by
2JooYeon
Loading…
2 of 4 tasks
fix(http): enforce streaming timeouts outside reqwest
anthropic
Anthropic router changes
documentation
Improvements or additions to documentation
model-gateway
Model gateway crate changes
needs-rebase
PR has merge conflicts that need to be resolved
openai
OpenAI router changes
#1858
opened Jun 30, 2026 by
jshanson7
Contributor
Loading…
feat(multimodal): optimize EPD encode routing
dependencies
Dependency updates
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
multimodal
Multimodal crate changes
protocols
Protocols crate changes
#1853
opened Jun 28, 2026 by
chenht2022
Contributor
•
Draft
feat(multimodal): add EPD encode routing
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
protocols
Protocols crate changes
python-bindings
Python bindings changes
#1852
opened Jun 28, 2026 by
chenht2022
Contributor
Loading…
[DO NOT MERGE] perf(multimodal): reduce video decode, Qwen preprocess, and TokenSpeed handoff overhead
dependencies
Dependency updates
documentation
Improvements or additions to documentation
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
multimodal
Multimodal crate changes
tests
Test changes
#1820
opened Jun 23, 2026 by
yechank-nvidia
Collaborator
Loading…
feat(multimodal): configurable tensor transport + vLLM SHM and video
dependencies
Dependency updates
documentation
Improvements or additions to documentation
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
needs-rebase
PR has merge conflicts that need to be resolved
protocols
Protocols crate changes
python-bindings
Python bindings changes
tests
Test changes
#1818
opened Jun 22, 2026 by
slin1237
Collaborator
Loading…
3 tasks done
feat(tool_parser,reasoning_parser): add Step-3.5 tool and reasoning parsers
reasoning-parser
Reasoning parser changes
tests
Test changes
tool-parser
Tool/function call parser changes
#1817
opened Jun 22, 2026 by
slin1237
Collaborator
Loading…
2 of 3 tasks
feat(reasoning_parser): add Mistral/Magistral reasoning parser
reasoning-parser
Reasoning parser changes
#1816
opened Jun 22, 2026 by
slin1237
Collaborator
Loading…
2 of 3 tasks
feat(tool_parser,reasoning_parser): add MiniMax-M3 tool and reasoning parsers
reasoning-parser
Reasoning parser changes
tests
Test changes
tool-parser
Tool/function call parser changes
#1815
opened Jun 22, 2026 by
slin1237
Collaborator
Loading…
2 of 3 tasks
[blocked] chore(deps): bincode 1.3->3.0 (smg-mesh) — 3.0.0 is a tombstone compile_error release
dependencies
Dependency updates
mesh
Mesh crate changes
Discover local HTTP model ids from /v1/models
model-gateway
Model gateway crate changes
openai
OpenAI router changes
#1793
opened Jun 19, 2026 by
SpencerGarnets
Loading…
perf(least_load): sub-linear selection via power-of-two-choices
enhancement
New feature or request
model-gateway
Model gateway crate changes
priority:high
High priority
stale
PR has been inactive for 14+ days
#1787
opened Jun 18, 2026 by
slin1237
Collaborator
Loading…
feat(grpc-servicer): add Prometheus /metrics sidecar
dependencies
Dependency updates
enhancement
New feature or request
grpc
gRPC client and router changes
metrics-consolidation
Epic: Engine Metrics Consolidation & PD Observability
tests
Test changes
#1782
opened Jun 18, 2026 by
slin1237
Collaborator
Loading…
fix(observability): include gRPC workers in GET /engine_metrics
enhancement
New feature or request
grpc
gRPC client and router changes
metrics-consolidation
Epic: Engine Metrics Consolidation & PD Observability
model-gateway
Model gateway crate changes
tests
Test changes
#1779
opened Jun 18, 2026 by
slin1237
Collaborator
Loading…
[gateway] WebSocket transport for the Responses API — up to 45% lower TTFT, 22–48% lower ITL p99 tail, 17–21% less gateway CPU
dependencies
Dependency updates
documentation
Improvements or additions to documentation
grpc
gRPC client and router changes
model-gateway
Model gateway crate changes
tests
Test changes
#1770
opened Jun 18, 2026 by
Venkat2811
Loading…
5 tasks done
feat(smg): worker-sync observability — drops, spec fallback, refused tombstones, drift
mesh
Mesh crate changes
model-gateway
Model gateway crate changes
#1716
opened Jun 13, 2026 by
CatherineSue
Member
Loading…
2 of 4 tasks
feat(smg): cluster-wide rate limiting via epoch-windowed mesh counters
mesh
Mesh crate changes
model-gateway
Model gateway crate changes
#1715
opened Jun 13, 2026 by
CatherineSue
Member
Loading…
2 of 4 tasks
refactor(service-discovery): migrate k8s watcher to informer (reflector + Store)
model-gateway
Model gateway crate changes
python-bindings
Python bindings changes
stale
PR has been inactive for 14+ days
#1688
opened Jun 11, 2026 by
key4ng
Collaborator
Loading…
3 of 4 tasks
feat(mesh): causally-stable tombstone GC driven by per-peer ack watermarks
mesh
Mesh crate changes
tests
Test changes
#1686
opened Jun 11, 2026 by
CatherineSue
Member
Loading…
2 of 4 tasks
fix(model_gateway): reclaim per-worker mutation lock for absent worker ids
model-gateway
Model gateway crate changes
stale
PR has been inactive for 14+ days
#1684
opened Jun 11, 2026 by
slin1237
Collaborator
Loading…
2 of 4 tasks
fix(reasoning-parser): hold back split think_end_token across streaming chunks
reasoning-parser
Reasoning parser changes
stale
PR has been inactive for 14+ days
#1678
opened Jun 11, 2026 by
slin1237
Collaborator
Loading…
2 of 4 tasks
Previous Next
ProTip!
Updated in the last three days: updated:>2026-07-01.