-
Notifications
You must be signed in to change notification settings - Fork 24
OpenJun 26, 2026
Overdue by 9 day(s)
•Due by June 26, 2026
•Last updated Planned work: harden the OpenAI-compatible audio endpoints (TTS + STT, MLX FFI and panic safety), new models (Mellum 2) and Gemma 4 video input, a ComputeBackend abstraction seam, MoE and dense decode performance gaps vs mlx-lm, disaggregated-serving expansion (/v1/completions, multi-node routing and failover), and native Windows + CUDA builds with packaging hardening.
72% complete
List view
0 of 6 selected 0 issues of 6 selected
perf(core): adaptive selector for the native paged-attention decode kernel
area:coremlxcel-core: MLX FFI, primitives, KV cache, layersmlxcel-core: MLX FFI, primitives, KV cache, layersarea:inferenceGeneration, sampling, decoding (incl. speculative, DRY)Generation, sampling, decoding (incl. speculative, DRY)priority:mediumMedium priorityMedium prioritytype:performancePerformance improvementsPerformance improvementsStatus: Open.#331 In lablup/mlxcel;perf(moe): backend-aware fused-MoE Dff cap (CUDA crossover) and dispatch heuristic
area:coremlxcel-core: MLX FFI, primitives, KV cache, layersmlxcel-core: MLX FFI, primitives, KV cache, layersarea:modelsModel architectures, weights, loading, metadataModel architectures, weights, loading, metadatapriority:mediumMedium priorityMedium prioritytype:performancePerformance improvementsPerformance improvementsStatus: Open.#330 In lablup/mlxcel;perf(nemotron-h): decode gap is MoE-block op-density (routed + shared expert), not SSM/attention
area:inferenceGeneration, sampling, decoding (incl. speculative, DRY)Generation, sampling, decoding (incl. speculative, DRY)area:modelsModel architectures, weights, loading, metadataModel architectures, weights, loading, metadataplatform:macosmacOS (Apple Silicon) specificmacOS (Apple Silicon) specificpriority:mediumMedium priorityMedium prioritytype:performancePerformance improvementsPerformance improvementsStatus: Open.#284 In lablup/mlxcel;feat: need a logo
area:docsUser and developer documentationUser and developer documentationhelp wantedExtra attention is neededExtra attention is neededpriority:mediumMedium priorityMedium prioritytype:enhancementNew features, capabilities, or significant additionsNew features, capabilities, or significant additionsStatus: Open.#59 In lablup/mlxcel;chore: harden packaging environment to enforce 4-eyes review on signed releases
priority:lowLow priorityLow prioritystatus:backlogIn the backlog, not yet readyIn the backlog, not yet readytype:choreMaintenance tasks (build, CI, etc.)Maintenance tasks (build, CI, etc.)Status: Open.#6 In lablup/mlxcel;ci(deps): bump dtolnay/rust-toolchain from 1.93.1 to 1.100.0
type:choreMaintenance tasks (build, CI, etc.)Maintenance tasks (build, CI, etc.)type:dependencyDependency updatesDependency updatesStatus: Open (in progress).