-
Notifications
You must be signed in to change notification settings - Fork 60
Pull requests: kkokosa/dotLLM
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
fix(cuda): widen byte-size products on the weight-upload path (#432)
#433
opened Aug 9, 2026 by
jamesburton
Loading…
fix(cpu/matmul): widen GEMV weight row offsets to long (#429)
#431
opened Aug 7, 2026 by
jamesburton
Loading…
perf: hoist F16 weight row conversion out of the token loop
#430
opened Aug 4, 2026 by
unsafePtr
Loading…
test(unit): report hardware-gated tests as skipped instead of passed (#421)
#428
opened Jul 31, 2026 by
jamesburton
Loading…
docs(server): drop the unimplemented lora_adapter request field
#427
opened Jul 31, 2026 by
jamesburton
Loading…
server(sse): stop sending Connection: keep-alive on the streaming endpoints
#426
opened Jul 31, 2026 by
jamesburton
Loading…
perf(cpu): balanced work partitioning across compute threads (#402)
#420
opened Jul 31, 2026 by
jamesburton
Loading…
perf(tokenizers): reuse one BPE merge queue per Encode call (#413)
#419
opened Jul 31, 2026 by
jamesburton
Loading…
feat(eval): shared perplexity harness with llama.cpp-comparable sliding-window mode
#418
opened Jul 30, 2026 by
jamesburton
Loading…
fix(tokenizers): pre-tokenize raw text before byte-encoding; add StarCoder/SmolLM pre-type family
#417
opened Jul 30, 2026 by
jamesburton
Loading…
fix(gguf): parse v2 counts/lengths as uint64 (spec-correct)
#365
opened Jun 30, 2026 by
jamesburton
Loading…
feat(cuda): optional arch-tiered PTX dispatch with compute_61 fallback
#333
opened Jun 19, 2026 by
jamesburton
Loading…
feat: Microsoft.Extensions.AI IChatClient adapter (DotLLM.Extensions.AI) (#327)
#328
opened Jun 15, 2026 by
jamesburton
Loading…
server: Anthropic-compatible Messages API (/v1/messages) (#325)
#326
opened Jun 15, 2026 by
jamesburton
Loading…
Feature - Metal GPU backend for macOS (Apple Silicon)
#324
opened Jun 14, 2026 by
rflechner
Loading…
perf(cpu/matmul): Q8_0 outer-product GEMM via AVX2 VNNI (VPDPBUSD-256) — unblock #61 register pressure
#323
opened Jun 9, 2026 by
jamesburton
•
Draft
models(architecture): add Gemma 4 family — Architecture enum + HF detection + forward wiring (Phase 8 Step 57)
#320
opened Jun 8, 2026 by
jamesburton
•
Draft
5 tasks done
models(transformer): wire Gemma 3 per-layer SW + attn/final softcaps + QPAS
#318
opened Jun 8, 2026 by
jamesburton
•
Draft
4 tasks done
models(safetensors): SmolLM3 + Gemma 3 architecture detection (rebased onto #314)
#317
opened Jun 8, 2026 by
jamesburton
•
Draft
4 tasks done
diagnostics(sae): SparseAutoencoder hook + top-K feature extraction
#316
opened Jun 8, 2026 by
jamesburton
•
Draft
kernels(cpu)(matmul): F32 outer-product tiled GEMM kernel for prefill
#315
opened Jun 8, 2026 by
jamesburton
Loading…
loader(safetensors): port MLA / Mixtral / Qwen-MoE loaders onto ISafetensorsTensorSource
#314
opened Jun 8, 2026 by
jamesburton
•
Draft
models(transformer-perf): direct-to-cache K/V projection via TryReserveSlot
#310
opened Jun 8, 2026 by
jamesburton
•
Draft
engine(kv-cache): IKvCache.TryReserveSlot — write-into-cache primitive
#309
opened Jun 8, 2026 by
jamesburton
Loading…
tokenizers: TryDecode(Span<char>) zero-allocation overload + IncrementalDetokenizer adoption
#308
opened Jun 8, 2026 by
jamesburton
Loading…
Previous Next
ProTip!
Updated in the last three days: updated:>2026-08-31.