Skip to content
View yuyuyUuuuuuuuuuuu's full-sized avatar

Block or report yuyuyUuuuuuuuuuuu

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse

Popular repositories Loading

  1. deepseek-harness-dsh-loading-history-hang-fix deepseek-harness-dsh-loading-history-hang-fix Public

    Fix for the DeepSeek Harness (DSH) web UI stuck on "Loading history…": JavaScriptCore/WebKit lossless-JSON validator bug (Safari, iPhone, iPad), WebSocket heartbeat drops, 13 patches with installer…

    JavaScript

  2. dgx-spark-gb10-vllm-qwen3.6-benchmarks dgx-spark-gb10-vllm-qwen3.6-benchmarks Public

    Qwen3.6-35B-A3B on one NVIDIA DGX Spark (GB10, sm_121) with vLLM: FP8 vs NVFP4, MTP and DFlash speculative decoding, Marlin vs FlashInfer b12x MoE kernels, vLLM 0.23 vs 0.25. Measured tok/s, compos…

    HTML

  3. glm-5.3-flash-dgx-spark-vllm-optimization glm-5.3-flash-dgx-spark-vllm-optimization Public

    GLM-5.3-Flash (320B MoE, EXL3 4bpw) on 2x NVIDIA DGX Spark (GB10) with vLLM, TP=2: single-stream decode 45-99 tok/s by workload, ~3,000 tok/s prefill (32k prompt), 2M-token fp8 KV cache. Optimizati…

    Python

  4. vllm-hybrid-kv-nvme-spill vllm-hybrid-kv-nvme-spill Public

    vLLM v1 KV connector that saves per-conversation resume points (MLA + Mamba/linear-attention state + draft window) to local NVMe and restores them instead of re-prefilling; TP=2, per-request TTL, r…

    Python