Commit Graph

50 Commits

Author SHA1 Message Date
Liangsheng Yin
fd9ad817ec Organize image inputs (#1531) 2024-09-29 06:28:55 +00:00
Lianmin Zheng
2854a5ea9f Fix the overhead due to penalizer in bench_latency (#1496) 2024-09-23 07:38:14 -07:00
Ke Bao
c6b6d2e71b Enable MLA by default (#1447) 2024-09-17 11:42:48 +00:00
Lianmin Zheng
2fa5cec775 Simplify sampler and its error handling (#1441) 2024-09-16 21:23:31 -07:00
Lianmin Zheng
9ba1f09760 [Fix] Fix logprob and normalized_logprob (#1428) 2024-09-15 06:36:06 -07:00
Liangsheng Yin
70b6802982 Optimize conflicts between CUDA graph and vocab mask tensors (#1392) 2024-09-13 20:27:53 -07:00
Lianmin Zheng
b912de11b0 Make stop reason a dict instead of str (#1407) 2024-09-12 20:47:31 -07:00
Ying Sheng
712216928f [Feature] Initial support for multi-LoRA serving (#1307) 2024-09-12 16:46:14 -07:00
Lianmin Zheng
fec185ce0c Refactor attention backend (#1381) 2024-09-11 11:44:26 -07:00
Liangsheng Yin
144bc70fcc Organize flashinfer indices update (#1378) 2024-09-10 17:38:59 -07:00
Lianmin Zheng
46094e0c1b Deprecate --disable-flashinfer and introduce --attention-backend (#1380) 2024-09-10 17:11:16 -07:00
Liangsheng Yin
fbb4754cb8 Fix vocab mask update bug (#1376) 2024-09-10 13:10:36 -07:00
Liangsheng Yin
69b3bb9ae1 Unify forward mode (#1360) 2024-09-09 13:49:29 -07:00
Kaichen Zhang - NTU
662ecd9368 [Feat] Add modalities for vision server when handling pixel values for llava (#1346) 2024-09-09 02:07:34 -07:00
Lianmin Zheng
1e495e0847 [Fix] Fix select by ensuring each request has at least one token (#1318) 2024-09-03 06:31:45 -07:00
Liangsheng Yin
381dd57bd6 Sampler cudagraph (#1253) 2024-08-28 18:58:52 -07:00
Lianmin Zheng
bf53bf5142 [Fix] Fix llava on multi images (#1247) 2024-08-28 06:33:05 -07:00
Yineng Zhang
f25f4dfde5 hotfix: revert sampler CUDA Graph (#1242) 2024-08-28 21:16:47 +10:00
havetc
909f34363b [FIX] Wrong logger (#1230) 2024-08-27 20:10:46 +10:00
havetc
9935f97b3e [FEAT] JSON constrained support (#1125)
Co-authored-by: Yineng Zhang <me@zhyncs.com>
2024-08-26 09:37:26 -07:00
Liangsheng Yin
75ce37f401 Move sampler into CUDA graph (#1201)
Co-authored-by: Yineng Zhang <me@zhyncs.com>
2024-08-26 07:02:50 -07:00
Lianmin Zheng
f6af3a6561 Cleanup readme, llava examples, usage examples and nccl init (#1194) 2024-08-24 08:02:23 -07:00
Liangsheng Yin
364d3d72a7 Fix broken penalty (#1184) 2024-08-22 08:16:35 +00:00
Liangsheng Yin
83e23c69b3 Improve code style of sampler (#1168) 2024-08-21 16:48:24 -07:00
intervitens
068e9eae55 Support min-p sampling (#1167) 2024-08-21 22:49:32 +00:00
Lianmin Zheng
bea2bb9eea Improve multi-node stability (#1171) 2024-08-20 22:35:05 -07:00
Lianmin Zheng
a8ae640328 Improve docs and warnings (#1164) 2024-08-20 08:31:29 -07:00
Liangsheng Yin
3694f8f996 Mixed style of chunked prefill (#1013) 2024-08-16 09:13:00 +00:00
Lianmin Zheng
5a261bd055 Fix the deadlock in multi-node tp (#1122) 2024-08-16 01:39:24 -07:00
Liangsheng Yin
73cf6834f2 Support stop_token_ids in sglang API (#1092) 2024-08-15 00:31:39 +00:00
Liangsheng Yin
a34dd86a7d Use dtype to control generate (#1082)
Co-authored-by: zhyncs <me@zhyncs.com>
2024-08-14 15:58:07 +00:00
Liangsheng Yin
7de6034534 Fix the prefix indices (#1037) 2024-08-11 17:57:02 -07:00
Liangsheng Yin
43fbb6d919 Fix input_ids && rename to fill_ids (#1021) 2024-08-10 16:24:12 -07:00
Liangsheng Yin
62757db6f0 Reduce the overhead when cache is disabled (#1010) 2024-08-09 16:36:57 -07:00
gryffindor-rr
9cf0a5bada Add skip_tokenizer_init args. (#959)
Co-authored-by: lzhang <zhanglei@modelbest.cn>
2024-08-09 12:14:13 -07:00
Juwan Yoo
10bca45bc6 bugfix: penalizers to be merged before reqs (#1001) 2024-08-09 21:46:24 +10:00
Ying Sheng
e040a2450b Add e5-mistral embedding model - step 3/3 (#988) 2024-08-08 16:31:19 -07:00
Juwan Yoo
ab7875941b feat: frequency, min_new_tokens, presence, and repetition penalties (#973) 2024-08-08 04:21:08 -07:00
yichuan~
3a79613c28 support more optioin about usage in stream mode (#985)
Co-authored-by: Ying Sheng <sqy1415@gmail.com>
2024-08-08 09:41:57 +00:00
Liangsheng Yin
1ac304eeb4 Adjust InputeMetadata and ScheduleBatch (#981) 2024-08-08 01:11:22 -07:00
Liangsheng Yin
2b8257f325 Adjust max prefix len (#980) 2024-08-08 00:41:26 +00:00
Liangsheng Yin
7623091d97 RadixCache method adjust (#977) 2024-08-07 15:52:24 -07:00
Zhiqiang Xie
6db27f7b3b misc: correct the int data type for token ids and indices (#969) 2024-08-08 04:40:07 +08:00
Liangsheng Yin
7fa54a1ab3 Make req_pool_indices on CPU (#960) 2024-08-07 01:41:25 -07:00
Liangsheng Yin
87e8c090e9 Organize code (rename, movement) (#953) 2024-08-06 20:50:32 -07:00
Ke Bao
e1eae1fd15 Support MLA for DeepSeek-V2 with Triton - step 1 (#905) 2024-08-05 03:40:33 +10:00
min-xu-et
7dd8a7e6d9 fixed an error handling in bench_latency.py (#904) 2024-08-03 17:42:17 -07:00
Liangsheng Yin
c020f9ceda Support chunked prefill when radix cache is disabled (#811) 2024-08-01 00:29:01 -07:00
Ying Sheng
e7487b08bc Adjust default mem fraction to avoid OOM (#823) 2024-07-30 01:58:31 -07:00
Liangsheng Yin
cdcbde5fc3 Code structure refactor (#807) 2024-07-29 23:04:48 -07:00