nextn

Here are 5 public repositories matching this topic...

llama.cpp fork with TurboQuant WHT-rotated KV cache & weight compression + Gemma 4 MTP and Qwen 3.6 NextN speculative decoding (+30-50% throughput).

Qwen3.8-Flash-Next as Cogni-Brain on NVIDIA DGX Spark (GB10): HashK GPU PLE + SGLang NEXTN, 36.8 tok/s code, 100/100 tool-eval, 262K context.

reasoning speculative-decoding sglang agentic-ai tool-calling nvfp4 dgx-spark grace-blackwell sm121 fp8-kv-cache nextn cogni-brain qwen38-flash-next hashk spark-brain

Extract and verify MTP/NextN GGUF speculative draft models

mtp llama-cpp gguf speculative-decoding nextn

Minimal pinned SGLang recipe for Qwen3.8-Flash-Next on one DGX Spark with exact-FP8 SSD-backed PLE, native 262K context, and NEXTN.

nvidia llm-inference qwen speculative-decoding sglang qwen3 nvfp4 dgx-spark nextn

Run 180B-parameter Qwen3.8-Flash-Next on a DGX Spark with full 262K context using SSD-backed memory for ultimate desktop AI performance.

benchmark nvidia aarch64 blackwell openai-api vllm llm-inference qwen speculative-decoding sglang qwen3 gb10 nvfp4 dgx-spark cuda-graphs nextn

To associate your repository with the nextn topic, visit your repo's landing page and select "manage topics."