$ git log --author="am17an" --all --oneline
Every morning, RepoJournal's newsroom reads what shipped across the open-source projects developers depend on — and writes the wire. Your commits kept making the news. This is your clippings file: the stories where your work was the story.
17
stories filed
15
daily wires
$ grep -rn "author: am17an" wires/ | sort -r
Local LLMs
wire
2026-11-18
Local LLMs
wire
2026-11-10
llama.cpp's RPC layer gets faster and stops aborting on bad nodes
- RPC: Improve loading times ggml-org/llama.cpp
Local LLMs
wire
2026-11-08
Metal GROUP_NORM fix, Ripple preview lands in Ollama's MLX runner
- llama_moe_cache: run missed experts on CPU, add tuning script ggml-org/llama.cpp
Local LLMs
wire
2026-11-03
vLLM 0.33.0 lands alongside three open advisories
- llama + CUDA: flash_attn_ext_rows (fix unified kv for multi-seq) ggml-org/llama.cpp
Local LLMs
wire
2026-10-30
Faster Metal MXFP4 decode, a GLM-4.7 streaming parser stall fixed
- metal : faster MXFP4 mat-vec, full threadgroups for unary/binary/GLU ggml-org/llama.cpp
Local LLMs
wire
2026-10-25
MoE expert masking lands across every llama.cpp backend
- model: DSV4.1 flash (#31086) ggml-org/llama.cpp
Local LLMs
wire
2026-10-18
vLLM v0.32.0 lands system-one endpoint; Ollama exposes template_source
- CUDA: faster MMVQ for low-bit types ggml-org/llama.cpp
Local LLMs
wire
2026-10-15
llama.cpp lands GDN chunked kernels and SYCL IQ3 reorder
- cuda, hip: GDN chunked kernel ggml-org/llama.cpp
Local LLMs
wire
2026-10-07
vLLM's multimodal input path is remote code execution, and MLX's sort is silently wrong
- RPC: add `-sm tensor` ggml-org/llama.cpp
Local LLMs
wire
2026-10-03
llama.cpp spec sampling goes probabilistic, vLLM trims GLM-5.3 attention overhead
- CUDA: fuse shared experts into MMVQ ggml-org/llama.cpp
Local LLMs
wire
2026-10-02
Qwen4Exp MTP lands in llama.cpp as SGLang ships v0.5.21
- Qwen4Exp: add MTP ggml-org/llama.cpp
- llama: fix qwen4exp ggml-org/llama.cpp
Local LLMs
wire
2026-09-21
llama.cpp sanitizes invalid UTF-8 in the parser AST
- CUDA: enable sparse fa for qwen4 ggml-org/llama.cpp
Local LLMs
wire
2026-08-31
Vulkan top-k radix sort, HIP Q2_0 optimizations, and vLLM CUDA-graph memory fix
- ggml: add SWIGLU_CLAMP ggml-org/llama.cpp
Local LLMs
wire
2026-08-24
VLLM'S P2P WEIGHT SYNC CUTS RL MOE TRANSFER COSTS
- DeepseekV4: fix rollback with multi-seq ggml-org/llama.cpp
Local LLMs
wire
2026-08-03
DEEPSEEK V4 INFERENCE GETS 50% SPEEDUP; LLAMA.CPP AND VLLM BOTH SHIPPING OPTIMIZED PATHS
- DeepseekV4 MTP + DSpark ggml-org/llama.cpp
- convert: add option to create separate dspark GGUF ggml-org/llama.cpp