Skip to content

Actions: InfiniTensor/InfiniLM

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
766 workflow runs
766 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

refactor(transformer-cpu): Transformer::decode 支持多 Request 同时计算 CI #116: Commit 408b05f pushed by YdrMaster
March 4, 2024 07:49 2m 40s main
main
March 4, 2024 07:49 2m 40s
refactor(transformer-cpu): Transformer::update 支持多 Request 同时计算 CI #115: Commit cb3c6ff pushed by YdrMaster
March 4, 2024 07:35 2m 35s main
main
March 4, 2024 07:35 2m 35s
refactor(transformer-cpu): 整理 kernels，简化调用 CI #114: Commit 66ec32e pushed by YdrMaster
March 4, 2024 06:06 2m 35s main
main
March 4, 2024 06:06 2m 35s
build(transformer-cpu): 每个 kernel 拆分到一个文件并修改 gather 使适应 batched tokens CI #113: Commit 8f50ee1 pushed by YdrMaster
March 4, 2024 05:52 2m 57s main
main
March 4, 2024 05:52 2m 57s
refactor(xtask): 简化构造服务的参数，添加一个简单的测试 CI #112: Commit fa38a38 pushed by YdrMaster
March 4, 2024 04:11 2m 38s main
main
March 4, 2024 04:11 2m 38s
refactor(xtask): 简化构造服务的参数，添加一个简单的测试 CI #111: Commit 8f45d11 pushed by YdrMaster
March 4, 2024 03:52 2m 42s main
main
March 4, 2024 03:52 2m 42s
refactor(xtask): 简化构造服务的参数，添加一个简单的测试 CI #110: Commit c7a1c54 pushed by YdrMaster
March 4, 2024 03:39 2m 37s main
main
March 4, 2024 03:39 2m 37s
feat(xtask): 暂时根据模型目录的名字推断需要对 prompt 做哪种预处理 CI #109: Commit 724b64c pushed by YdrMaster
March 4, 2024 03:01 3m 4s main
main
March 4, 2024 03:01 3m 4s
build: 服务基于 xtask 搭建，放开 Cargo.lock CI #108: Commit b6f6f51 pushed by YdrMaster
March 4, 2024 01:39 2m 36s main
main
March 4, 2024 01:39 2m 36s
feat(xtask): 搭建服务的基本结构 CI #107: Commit e82a387 pushed by YdrMaster
March 1, 2024 10:00 2m 37s main
main
March 1, 2024 10:00 2m 37s
refactor(xtask): 重构服务的结构 CI #106: Commit f2a60e5 pushed by YdrMaster
March 1, 2024 09:35 20s main
main
March 1, 2024 09:35 20s
feat(xtask): 搭建服务的基本结构 CI #105: Commit 78c8001 pushed by YdrMaster
March 1, 2024 09:11 2m 48s main
main
March 1, 2024 09:11 2m 48s
perf(transformer): 跳过无意义的 reform CI #104: Commit 0c2f76b pushed by YdrMaster
March 1, 2024 06:03 2m 34s main
main
March 1, 2024 06:03 2m 34s
refactor(tensor): reform 直接拷贝只需要二者非 1 维度 stride 相同 CI #103: Commit 9aa5ce5 pushed by YdrMaster
March 1, 2024 04:09 2m 35s main
main
March 1, 2024 04:09 2m 35s
refactor(tensor): crate 项目结构压平 CI #102: Commit 57a9265 pushed by YdrMaster
March 1, 2024 03:19 2m 36s main
main
March 1, 2024 03:19 2m 36s
refactor(tensor): 整理 Tensor transform，现在只有 Split 需要 Physical: Clone CI #101: Commit af02063 pushed by YdrMaster
March 1, 2024 02:54 2m 37s main
main
March 1, 2024 02:54 2m 37s
refactor(transformer-nvidia): 更新 cuda crate，显存抽象移动到上游 CI #100: Commit a436a88 pushed by YdrMaster
March 1, 2024 01:28 1m 2s main
main
March 1, 2024 01:28 1m 2s
feat(xtask): 通过参数控制是否添加 bos CI #99: Commit cff122d pushed by YdrMaster
February 29, 2024 09:13 1m 1s main
main
February 29, 2024 09:13 1m 1s
fix(xtask): 拷贝 vocabs.txt CI #98: Commit fa62f1a pushed by YdrMaster
February 29, 2024 07:24 1m 6s main
main
February 29, 2024 07:24 1m 6s
fix(xtask): cast 不改变类型也不能跳过 CI #97: Commit b959567 pushed by YdrMaster
February 29, 2024 07:18 1m 1s main
main
February 29, 2024 07:18 1m 1s
feat(xtask): 支持不同种类的 tokenizer CI #96: Commit e68deba pushed by YdrMaster
February 29, 2024 06:32 1m 10s main
main
February 29, 2024 06:32 1m 10s
feat(transformer-nvidia): 实现 cuda 端到端推理 CI #95: Commit fa272f8 pushed by YdrMaster
February 29, 2024 04:40 2m 36s main
main
February 29, 2024 04:40 2m 36s
feat(transformer-nvidia): 实现 swiglu CI #94: Commit 306fa62 pushed by YdrMaster
February 29, 2024 04:14 3m 4s main
main
February 29, 2024 04:14 3m 4s
feat(transformer-nvidia): 实现 softmax / attention CI #93: Commit 8f016e4 pushed by YdrMaster
February 29, 2024 03:17 2m 45s main
main
February 29, 2024 03:17 2m 45s
feat(transformer-nvidia): 添加 kv cache，实现 reform CI #92: Commit b91ab28 pushed by YdrMaster
February 29, 2024 01:16 2m 48s main
main
February 29, 2024 01:16 2m 48s
ProTip! You can narrow down the results and go further in time using created:<2024-02-29 or the other filters available.