Commit Graph

1545 Commits

Author SHA1 Message Date
Liangsheng Yin
e7ebecf82e Fix cache hit rate when chunked prefill (#2555) 2024-12-26 03:14:28 -08:00
Xiaoyu Zhang
9a23c48456 h100 tuning fused_moe_triton for qwen2 moe (#2560) 2024-12-26 03:13:31 -08:00
Yineng Zhang
635a042623 docs: update deepseek v3 example (#2592) 2024-12-26 17:43:37 +08:00
Yineng Zhang
2dccecf432 fix: only enable moe_align_block_size for now (#2590) 2024-12-26 16:56:59 +08:00
Yineng Zhang
75ad0a143f docs: add deepseek v3 launch instructions (#2589) 2024-12-25 23:26:54 -08:00
Yineng Zhang
efc52f85e2 chore: bump v0.4.1 (#2582) 2024-12-26 07:14:51 +08:00
Yineng Zhang
60e2fdcf4f use sgl-kernel moe_align_block_size (#2581)
Co-authored-by: ispobock <ispobaoke@163.com>
Co-authored-by: HandH1998 <1335248067@qq.com>
2024-12-26 06:29:08 +08:00
Yineng Zhang
d7c0e872b0 chore: bump 0.0.2.post8 for sgl-kernel (#2580) 2024-12-26 06:11:39 +08:00
Yineng Zhang
31548116a8 fix moe_align_block_size_kernel for shared memory issue (#2579)
Co-authored-by: ispobock <ispobaoke@163.com>
2024-12-26 05:31:04 +08:00
HandH1998
53aed988cb Refactor MoE (#2575)
Co-authored-by: zhyncs <me@zhyncs.com>
2024-12-26 00:02:14 +08:00
Ying Sheng
8a56b43175 [Bench] Flush cache before benchmarking (#2566) 2024-12-24 11:21:21 +08:00
Ke Bao
e835a50021 Reorg moe code (#2563) 2024-12-24 01:10:22 +08:00
Lianmin Zheng
23e5e50fd5 Fix gemlite import (#2553) 2024-12-22 20:21:17 -08:00
Shi Shuai
25e5d589e3 Doc: Update Grammar Backend (#2545)
Co-authored-by: Chayenne <zhaochen20@outlook.com>
2024-12-22 17:14:40 -08:00
Lianmin Zheng
41b1db69b8 A better aio rwlock that guarantees the order (#2547) 2024-12-22 15:44:32 -08:00
Lianmin Zheng
8496701934 [Misc] Fix metrics, weight update lock, request logging (#2543) 2024-12-22 06:27:22 -08:00
Xiaoyu Zhang
7d672d277b [kernel optimize] benchmark write_req_to_token_pool_triton and optimize kernel (#2509) 2024-12-22 02:31:02 -08:00
Yineng Zhang
d4b174817d docs: update sponsorship (DataCrunch) (#2523) 2024-12-22 02:29:04 -08:00
Lei
19ba2b0ea9 Add lora_paths to v1_chat_generate_request (#2529) 2024-12-22 02:23:33 -08:00
Yineng Zhang
4e1e3cff20 fix #2528 (#2541) 2024-12-22 00:14:41 +08:00
Yineng Zhang
8f4d04e540 chore: bump v0.4.0.post2 (#2525) 2024-12-21 21:16:34 +08:00
Jerry Zhang
feb2b768ba Add integration with gemlite weight only quant (#2528) 2024-12-21 00:25:25 +08:00
Yineng Zhang
d95a5f5bf5 fix followup #2517 (#2524) 2024-12-19 23:24:30 +08:00
Yineng Zhang
4b83db24f1 fix: continue to use flashinfer 0.1.6 temporarily (#2517) 2024-12-19 14:03:24 +08:00
Yineng Zhang
64456cf023 docs: update README (#2516) 2024-12-19 13:44:02 +08:00
Yineng Zhang
bb4a922023 feat: add llama3 eval (#2515) 2024-12-19 13:37:09 +08:00
Lianmin Zheng
21e9e63ad5 Print progress bar during cuda graph capture (#2502) 2024-12-17 06:33:46 -08:00
Lianmin Zheng
1fc84cf60b Update readme (#2500)
Co-authored-by: Ravi Theja <ravi03071991@gmail.com>
Co-authored-by: “yixin-huang1” <yixinhuang1@berkeley.edu>
2024-12-17 04:33:36 -08:00
Lianmin Zheng
361ea8d912 Fix openai protocols and pass top_k, min_p (#2499) 2024-12-17 04:14:14 -08:00
Lei
33c5ff2845 Add lora_path to chat completion (#2438) 2024-12-17 03:47:49 -08:00
Hui Liu
5ce9daea59 ROCm support for sglang.check_env (#2426) 2024-12-17 03:45:14 -08:00
Ata Fatahi
ce094a5d79 Clean up GPU memory after killing sglang processes (#2457)
Signed-off-by: Ata Fatahi <immrata@gmail.com>
2024-12-17 03:42:40 -08:00
bjmsong
e21026690d benchmark decoding attention kernel with cudnn (#2467)
Co-authored-by: root <bjmsong@126.com>
2024-12-17 03:31:57 -08:00
Lianmin Zheng
bd6196163e Small fix for the order of apply_torchao_config (#2495) 2024-12-16 19:21:11 -08:00
Lianmin Zheng
56198b45d9 Add a benchmark script for in-batch prefix caching (#2494) 2024-12-16 18:49:02 -08:00
Lianmin Zheng
ba36b5520a Revert "Small fixes for torchao quant" (#2493) 2024-12-16 15:04:16 -08:00
Lianmin Zheng
9cd9dc83b3 Temporarily disable unit test of torch native attention backend (#2492) 2024-12-16 14:17:27 -08:00
Lianmin Zheng
7a1aecb938 Simplify pytorch sampling kernel and logit processor (#2491) 2024-12-16 14:11:09 -08:00
Jerry Zhang
82699474fd Small fixes for torchao quant (#2476) 2024-12-16 14:08:12 -08:00
Yineng Zhang
7154b4b1df minor: update flashinfer nightly (#2490) 2024-12-16 23:02:49 +08:00
xiaobochen
b532a5fd16 fix moe-ep accuracy issue for fp8 (#2489) 2024-12-16 20:54:02 +08:00
Xiaoyu Zhang
a0592c059f [Benchmark] add a benchmark for hf/vllm/sglang rmsnorm (#2486) 2024-12-15 13:52:08 +08:00
Yineng Zhang
e8dbdf75bc fix typo (#2487) 2024-12-15 13:44:55 +08:00
yizhang2077
e04d3f2897 adapt tensorrt llm custom all reduce to sgl-kernel (#2481)
Co-authored-by: Yineng Zhang <me@zhyncs.com>
2024-12-15 13:15:59 +08:00
Yineng Zhang
5f2595be43 hotfix: checking for HIP (#2485) 2024-12-15 02:47:26 +08:00
Ke Bao
0ba2c58947 Remove cuda graph batch size adjustment for dp attention (#2484) 2024-12-14 23:53:54 +08:00
Yineng Zhang
fccbfa3752 format: add clang-format for sgl-kernel (#2483) 2024-12-14 22:36:04 +08:00
Ke Bao
2f9bd0fafd Fix correctness issue for triton decoding kernel (#2479) 2024-12-14 16:50:54 +08:00
Lianmin Zheng
5282a4735f [Minor] Fix grok model loader (#2473) 2024-12-12 14:34:47 -08:00
Yineng Zhang
f0ed9c353e feat: support dev image (#2469) 2024-12-13 02:23:52 +08:00