gitlawr
|
b4bc99802c
|
feat: add benchmark
|
10 months ago |
|
gitlawr
|
291ff064f0
|
fix: update size test
|
10 months ago |
thxCode
|
5e8a0b561a
|
feat: support mindie
Signed-off-by: thxCode <thxcode0824@gmail.com>
|
10 months ago |
|
gitlawr
|
d1e8dc22d0
|
chore: update exception message for empty architectures
|
11 months ago |
|
gitlawr
|
887f41e9c4
|
chore: bump box
|
11 months ago |
|
gitlawr
|
59bf544908
|
chore: update vram claim env
|
11 months ago |
|
gitlawr
|
db4b780b7c
|
feat: add evaluation cache
|
11 months ago |
|
gitlawr
|
0a5facdb40
|
fix: update non-LLM vllm claim
|
11 months ago |
|
gitlawr
|
95916b82b5
|
feat: evaluated resource claim
|
11 months ago |
|
gitlawr
|
592723e291
|
chore: update cosyvoice models
|
11 months ago |
|
gitlawr
|
d50e2b95f3
|
feat: add model info in profiling
|
11 months ago |
|
gitlawr
|
f01ee2e267
|
fix: migrate legacy hf cache
|
11 months ago |
|
gitlawr
|
cb8d75c4e0
|
chore: update box
|
11 months ago |
|
gitlawr
|
32721e754c
|
feat: add deepseek-v3-0324
|
11 months ago |
|
gitlawr
|
00e53c9a00
|
chore: update evaluate message
|
11 months ago |
|
gitlawr
|
ef3cbac12e
|
fix: use soft filelock
|
11 months ago |
|
gitlawr
|
69c6fac56f
|
feat: validation for dist vllm limit per worker
|
11 months ago |
|
gitlawr
|
5f312e9ed3
|
feat: add model evaluations
|
11 months ago |
|
gitlawr
|
160fefa8de
|
refactor: rename claim dataclass
|
11 months ago |
|
gitlawr
|
1f20e208b8
|
chore: update delaying restart log level
|
11 months ago |
|
gitlawr
|
d85a861eee
|
chore: bump dependencies
|
11 months ago |
|
gitlawr
|
25930f2075
|
refactor: use local dir in hf downloader
|
11 months ago |
|
gitlawr
|
f386508e13
|
refactor: add tini
|
11 months ago |
|
gitlawr
|
80ad177a21
|
Revert "fix: In the terminate_process method, adding wait after a failed kill operation helps reclaim the child process and prevents the creation of zombie processes"
This reverts commit 8796ca44b4.
|
11 months ago |
|
gitlawr
|
e72c0a45ab
|
fix: resolved paths contains
|
11 months ago |
|
gitlawr
|
ecc45a3b69
|
fix: restart ray on exit
|
11 months ago |
|
gitlawr
|
7e700545c5
|
chore: bump dependencies
|
11 months ago |
|
gitlawr
|
0d2423e0bb
|
fix: none file exception
|
11 months ago |
|
gitlawr
|
e30747e66c
|
fix: bool gpustack env
|
11 months ago |
|
gitlawr
|
4e5e954803
|
fix: compute allocatable vram
|
11 months ago |
|
gitlawr
|
ffedf70089
|
refactor: move gpu filtering from worker filter to gguf resource selector
|
11 months ago |
|
gitlawr
|
1fefb68a34
|
feat: get sharded file paths for localpath files
|
11 months ago |
|
gitlawr
|
699e863325
|
fix: remove instances on worker deletion
|
11 months ago |
|
gitlawr
|
ed47e61939
|
feat: support restart on errors
|
11 months ago |
|
gitlawr
|
b0d7d08bdf
|
fix: powershell lint
|
11 months ago |
|
gitlawr
|
7af01766a1
|
fix: empty vram claim
|
11 months ago |
|
gitlawr
|
d1efdf5730
|
fix: match existing model files for local path models
|
11 months ago |
|
gitlawr
|
8c0e3ef8c0
|
fix: reverse local verify
|
11 months ago |
|
gitlawr
|
6f05dd9f0f
|
feat: detect reranker architectures
|
11 months ago |
|
gitlawr
|
a0df56059d
|
fix: missing token usage in vllm reranker
|
11 months ago |
|
gitlawr
|
f6451f1ef4
|
fix: avoid repatching tqdm
|
11 months ago |
|
gitlawr
|
56e618d1be
|
fix: no file match on empty filename
|
11 months ago |
|
gitlawr
|
b90f88fb59
|
feat: cleanup option on model file deletion
|
11 months ago |
|
gitlawr
|
a1638344a7
|
fix: download task exception
|
11 months ago |
|
gitlawr
|
9409565daf
|
fix: get config attr
|
11 months ago |
|
gitlawr
|
164375f4a7
|
feat: add worker name label
|
11 months ago |
Jiangzhou Li
|
8796ca44b4
|
fix: In the terminate_process method, adding wait after a failed kill operation helps reclaim the child process and prevents the creation of zombie processes
|
11 months ago |
|
Jiangzhou Li
|
edfbd61c6f
|
feat: Add an option to use hf_transfer for HF Hub downloads
fix: update comment
fix: update start.md
|
11 months ago |
|
gitlawr
|
47412f9cf7
|
fix: docker ci
|
11 months ago |
|
Jiangzhou Li
|
d9f409ee4f
|
fix: avoid rpc server and inference server use same name
fix code format
fix: code format error
fix: windows also try symlink
fix: use hardlink in windows
fix: Optimize code format
fix: resolve CR question
|
11 months ago |
