Skip to content

Actions: ztxz16/fastllm

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
465 workflow runs
465 workflow runs

Filter by Workflow

⌥ + click/return to exclude

Filter by Event

⌥ + click/return to exclude

Filter by Status

⌥ + click/return to exclude

Filter by Branch

⌥ + click/return to exclude

Filter by Actor

⌥ + click/return to exclude
优化MLA解码调度计划复用 Action Build #766: Commit 05284b4 pushed by ztxz16
7m 13s master
master
7m 13s
优化单token路由与小维度GEMV Action Build #765: Commit a408800 pushed by ztxz16
5m 35s master
master
5m 35s
修复 Naive 解码测试的 CUDA Graph 编译 Action Build #764: Commit fc5b0f9 pushed by ztxz16
6m 47s master
master
6m 47s
接入 FlashInfer 加速 GLM 稀疏预填充 Action Build #763: Commit 7187745 pushed by ztxz16
7m 6s master
master
7m 6s
优化 Naive Indexer 的 TopK 选择 Action Build #762: Commit a746576 pushed by ztxz16
7m 7s naive-topk-select-model-20261003
naive-topk-select-model-20261003
7m 7s
合并 Naive 推理优化 Action Build #761: Commit 42bfc71 pushed by ztxz16
5m 31s master
master
5m 31s
补充 MLA 精度与性能验证结果 Action Build #760: Commit b0a5a60 pushed by ztxz16
7m 3s master
master
7m 3s
补充 MLA 精度与性能验证结果 Action Build #759: Commit b0a5a60 pushed by ztxz16
7m 5s glm53-nvfp4-compact-20261003
glm53-nvfp4-compact-20261003
7m 5s
优化 Naive MoE 预填充调度 Action Build #758: Commit ec92bd0 pushed by ztxz16
7m 1s naive-moe-prefill-20261003
naive-moe-prefill-20261003
7m 1s
优化 Naive 索引器预填充 Action Build #757: Commit 6be7423 pushed by ztxz16
7m 23s naive-indexer-mma-20261002
naive-indexer-mma-20261002
7m 23s
补充 GGUF 优化与整理实测文档 Action Build #756: Commit 2a3c71c pushed by ztxz16
6m 47s master
master
6m 47s
优化 Naive prefill 并补充实测 Action Build #755: Commit 2faae1d pushed by ztxz16
7m 7s master
master
7m 7s
修复 Draft 验证脚本动态模块注册 Action Build #754: Commit 2700707 pushed by ztxz16
6m 30s master
master
6m 30s
整理 Flash-Next 用户手册与首页入口 Action Build #753: Commit 22d9c42 pushed by ztxz16
4m 34s master
master
4m 34s
优化 DeepSeek TP 无图推测校验调度 Action Build #752: Commit a2bf07f pushed by ztxz16
4m 20s master
master
4m 20s
移除 DFlash 实验环境开关 Action Build #751: Commit 90fc11e pushed by ztxz16
6m 47s master
master
6m 47s
修复工具调用流结束处理 Action Build #750: Commit 19909db pushed by ztxz16
6m 41s master
master
6m 41s
补充双卡三十二图完整任务验证 Action Build #749: Commit 3907878 pushed by ztxz16
6m 48s master
master
6m 48s
修复投机解码容量边界重复预填充 Action Build #748: Commit de8b299 pushed by ztxz16
6m 50s master
master
6m 50s
复用 DFlash 草稿计算缓冲并捕获 CUDA Graph Action Build #747: Commit 0027621 pushed by ztxz16
6m 41s master
master
6m 41s
精简 Qwen 专家选择流程 Action Build #746: Commit 4f344cd pushed by ztxz16
6m 43s master
master
6m 43s
扩展V100单卡长上下文KV直读 Action Build #745: Commit 629fac9 pushed by ztxz16
5m 27s master
master
5m 27s
默认量化兼容的NVFP4草稿 Action Build #744: Commit 7e0d670 pushed by ztxz16
6m 33s master
master
6m 33s
修复 CUDA Graph 临时内存分配与回退 Action Build #743: Commit 82ebf07 pushed by ztxz16
6m 40s master
master
6m 40s
修复 Qwen TP 默认设备映射 Action Build #742: Commit 37fcad2 pushed by ztxz16
6m 36s master
master
6m 36s