Search results

[Feature] Support user-selectable LoRA adapter files at load/runtime

architecture

enhancementNew feature or requestNew feature or request

Need Review

Status: Open.

#37 In nntrainer/Quick.AI;

· dlwlzzero opened on Jun 5, 2026

[Feat(API)] Accept full accumulated context and compute only the unprocessed delta via recorded context length

api

enhancementNew feature or requestNew feature or request

Status: Open.

#34 In nntrainer/Quick.AI;

· dlwlzzero opened on Jun 4, 2026

[Feasibility] Unify the nntrainer KV-cache manager with the QNN KV-cache manager

architecture

Need Review

Status: Open.

#33 In nntrainer/Quick.AI;

· dlwlzzero opened on Jun 4, 2026

[Refactor] Move the qnn/ context library into the nntrainer submodule

Status: Open.

#32 In nntrainer/Quick.AI;

· dlwlzzero opened on Jun 4, 2026

[Feat] Support concurrent multi-session inference

api

architecture

enhancementNew feature or requestNew feature or request

Status: Open.

#31 In nntrainer/Quick.AI;

· dlwlzzero opened on Jun 4, 2026

[Feat] Expose per-step logit values through the model-run API

api

enhancementNew feature or requestNew feature or request

Status: Open.

#30 In nntrainer/Quick.AI;

· dlwlzzero opened on Jun 4, 2026

[model] Support SLM

Status: Open.

#26 In nntrainer/Quick.AI;

· Seunghui98 opened on May 11, 2026

[nntrainer/PR] port #3850 from upstream nntrainer

Status: Open.

#19 In nntrainer/Quick.AI;

· EunjuYang opened on Apr 29, 2026

[Issue] 30B MoE with FSU (slimMoE, slimCachedMoE) fails with mmap-read=true option

Status: Open.

#13 In nntrainer/Quick.AI;

· EunjuYang opened on Apr 28, 2026

[model] Support gliner2-multi-v1 model

Status: Open.

#4 In nntrainer/Quick.AI;

· Seunghui98 opened on Apr 24, 2026

[model] Support tiny-BERT

Status: Open.

#3 In nntrainer/Quick.AI;

· Seunghui98 opened on Apr 24, 2026