Skip to content

Actions: TylunasLi/fastllm

Actions

All workflows

Actions

Loading...
Loading

Showing runs from all workflows
146 workflow runs
146 workflow runs

Filter by Event

Filter by Status

Filter by Branch

Filter by Actor

修复Android编译 Action Build #109: Commit 8e2466e pushed by TylunasLi
June 10, 2024 14:34 1m 29s win64
win64
June 10, 2024 14:34 1m 29s
修复Android编译 Action Build #108: Commit 580c302 pushed by TylunasLi
June 10, 2024 08:43 1m 35s win64
win64
June 10, 2024 08:43 1m 35s
修复Android编译 Action Build #107: Commit 2467ba9 pushed by TylunasLi
June 10, 2024 05:39 2m 15s win64
win64
June 10, 2024 05:39 2m 15s
修复Windows编译 Action Build #106: Commit 9bc47eb pushed by TylunasLi
June 9, 2024 15:07 2m 0s win64
win64
June 9, 2024 15:07 2m 0s
Merge pull request #444 from TylunasLi/no_tensor_core Action Build #105: Commit f3cfc63 pushed by TylunasLi
May 30, 2024 15:36 2m 54s v0.9.999
v0.9.999
May 30, 2024 15:36 2m 54s
修复llama模型批量推理不一致问题 Action Build #104: Commit e8c80c9 pushed by TylunasLi
April 24, 2024 15:22 7m 19s llama
llama
April 24, 2024 15:22 7m 19s
加入int4下linear的向量化访存，提高推理速度。 Action Build #103: Commit da16e05 pushed by TylunasLi
April 24, 2024 15:17 2m 52s no_tensor_core
no_tensor_core
April 24, 2024 15:17 2m 52s
Merge pull request #444 from TylunasLi/no_tensor_core Action Build #102: Commit f3cfc63 pushed by TylunasLi
April 24, 2024 14:35 3m 1s master
master
April 24, 2024 14:35 3m 1s
增加int4 GEMV算子的向量化访存机制，优化单条推理速度 Action Build #101: Commit f7c3960 pushed by TylunasLi
April 24, 2024 00:17 3m 4s no_tensor_core
no_tensor_core
April 24, 2024 00:17 3m 4s
修复增加fp16算子后，在Maxwell等老平台的编译问题 Action Build #100: Commit 9ecf89b pushed by TylunasLi
April 11, 2024 17:12 2m 47s no_tensor_core
no_tensor_core
April 11, 2024 17:12 2m 47s
Merge pull request #442 from TylunasLi/llama Action Build #99: Commit d3dfc0a pushed by TylunasLi
April 7, 2024 15:34 2m 47s master
master
April 7, 2024 15:34 2m 47s
加入int4下linear的向量化访存，提高推理速度。 Action Build #98: Commit fb4242e pushed by TylunasLi
March 31, 2024 05:53 2m 40s no_tensor_core
no_tensor_core
March 31, 2024 05:53 2m 40s
Merge pull request #441 from jacques-chen/master Action Build #97: Commit bf8f25f pushed by TylunasLi
March 31, 2024 05:48 2m 42s master
master
March 31, 2024 05:48 2m 42s
对齐同一模型python Tokenizer、Huggingface Tokenizers和SentencePiece的空格转换逻辑 Action Build #96: Commit 57b03e7 pushed by TylunasLi
March 30, 2024 15:42 3m 3s llama
llama
March 30, 2024 15:42 3m 3s
修复internLM2转换和推理的问题 Action Build #95: Commit 9affcbb pushed by TylunasLi
March 28, 2024 16:37 2m 51s llama
llama
March 28, 2024 16:37 2m 51s
Merge remote-tracking branch 'origin/master' into llama Action Build #94: Commit ef0b7a6 pushed by TylunasLi
March 23, 2024 04:13 2m 52s llama
llama
March 23, 2024 04:13 2m 52s
Merge pull request #439 from TylunasLi/win64 Action Build #93: Commit 0f22a2c pushed by TylunasLi
March 23, 2024 04:04 2m 58s master
master
March 23, 2024 04:04 2m 58s
Merge remote-tracking branch 'origin/master' into llama Action Build #92: Commit 285d3ad pushed by TylunasLi
March 23, 2024 03:56 2m 43s llama
llama
March 23, 2024 03:56 2m 43s
修复Cmake Windows编译下，int4乱码的问题 Action Build #91: Commit 32d1c6e pushed by TylunasLi
March 23, 2024 02:46 2m 42s win64
win64
March 23, 2024 02:46 2m 42s
修复Windows编译 Action Build #90: Commit 2f9dea2 pushed by TylunasLi
March 21, 2024 17:00 2m 47s win64
win64
March 21, 2024 17:00 2m 47s
Merge pull request #428 from TylunasLi/minicpm Action Build #89: Commit a0f2212 pushed by TylunasLi
March 21, 2024 16:56 2m 47s win64
win64
March 21, 2024 16:56 2m 47s
修复qwen1.5模型from_hf直接转换 Action Build #88: Commit 580a31b pushed by TylunasLi
March 21, 2024 15:14 3m 18s master
master
March 21, 2024 15:14 3m 18s
支持InternLM2的推理和转换 Action Build #87: Commit 71a7856 pushed by TylunasLi
March 20, 2024 16:14 4m 14s llama
llama
March 20, 2024 16:14 4m 14s
修复动态NTK外推计算，以及分组查询注意力下合并KQV的切分 Action Build #86: Commit bcf5432 pushed by TylunasLi
March 20, 2024 15:42 2m 42s llama
llama
March 20, 2024 15:42 2m 42s
优化MiniCPM算子，提速大约0.4% Action Build #85: Commit 7373729 pushed by TylunasLi
March 19, 2024 16:56 2m 42s minicpm
minicpm
March 19, 2024 16:56 2m 42s