Commits · 1fb76ebb93894447d1e3a0a04bca4875563b328c · change / sglang · GitLab

06 Jun, 2025 1 commit
- AITER backend extension and workload optimizations (#6838) · b819381f
  HAI authored Jun 05, 2025
```
Co-authored-by: wunhuang <wunhuang@amd.com>
Co-authored-by: Hubert Lu <Hubert.Lu@amd.com>
```
  b819381f
05 Jun, 2025 5 commits
- [CPU] enable CI for PRs, add Dockerfile and auto build task (#6458) · 562f279a
  Zaili Wang authored Jun 06, 2025
```
Co-authored-by: diwei sun <diwei.sun@intel.com>
Co-authored-by: Yineng Zhang <me@zhyncs.com>
```
  562f279a
- bugfix(OAI): Fix image_data processing for jinja chat templates (#6877) · 8b247489
  Chang Su authored Jun 05, 2025
  
  8b247489
- Support layerwise rebalancing experts (#6851) · 0de5e7d4
  fzyzcjy authored Jun 05, 2025
  
  0de5e7d4
- support 1 shot allreduce in 1-node and 2-node using mscclpp (#6277) · 8e3797be
  zyksir authored Jun 05, 2025
  
  8e3797be
- Support LoRA in TestOpenAIVisionServer and fix fused kv_proj loading bug. (#6861) · 4474eaf5
  Lifu Huang authored Jun 04, 2025
  
  4474eaf5
04 Jun, 2025 4 commits
- feat: add dp-rank to KV events (#6852) · f0f84975
  ishandhanani authored Jun 04, 2025
  
  f0f84975
- Decoder-only Scoring API (#6460) · 3f1e4339
  Chanh Nguyen authored Jun 04, 2025
```
Co-authored-by: Chanh Nguyen <cnguyen@linkedin.com>
```
  3f1e4339
- [Refactor] Multimodal data processing for VLM (#6659) · cf9815ba
  Xinyuan Tong authored Jun 04, 2025
```
Signed-off-by: Xinyuan Tong <justinning0323@outlook.com>
```
  cf9815ba
- [FEAT] Add transformers backend support (#5929) · 37f15475
  Marc Sun authored Jun 04, 2025
  
  37f15475
03 Jun, 2025 1 commit
- Add CPU optimized kernels for topk and rope fusions (#6456) · ff00895c
  jianan-gu authored Jun 03, 2025
  
  ff00895c
02 Jun, 2025 1 commit
- Add draft extend CUDA graph for flashinfer backend (#6805) · a2cb5913
  Ke Bao authored Jun 02, 2025
  
  a2cb5913
01 Jun, 2025 1 commit
- Update CI tests for Llama4 models (#6421) · c6a0cacc
  Ravi Theja authored Jun 01, 2025
  
  c6a0cacc
31 May, 2025 5 commits
- Improve profiler and integrate profiler in bench_one_batch_server (#6787) · 2d72fc47
  Lianmin Zheng authored May 31, 2025
  
  2d72fc47
- ci: relax test_function_call_required (#6786) · e39bca07
  Chang Su authored May 30, 2025
  
  e39bca07
- Temporarily lower mmlu threshold for triton sliding window backend (#6785) · a2bb8565
  Jianan Ji authored May 30, 2025
  
  a2bb8565
- feat(tool call): Enhance Llama32Detector for improved JSON parsing in non-stream (#6784) · f18b068f
  Chang Su authored May 30, 2025
  
  f18b068f
- update llama4 chat template and pythonic parser (#6679) · 4fac524b
  Chao Yang authored May 30, 2025
```
Co-authored-by: Chang Su <chang.s.su@oracle.com>
```
  4fac524b
30 May, 2025 1 commit
- Support sliding window in triton backend (#6509) · 22630ca2
  Jianan Ji authored May 30, 2025
  
  22630ca2
29 May, 2025 2 commits
- refactor(tool call): Fix BaseFormatDetector tool_index issue and refactor... · c673727e
  Chang Su authored May 29, 2025
```
refactor(tool call): Fix BaseFormatDetector tool_index issue and refactor `parse_streaming_increment` (#6715)
```
  c673727e
- Fix PP for Qwen3 MoE (#6709) · e06b0761
  iLeGend authored May 29, 2025
  
  e06b0761
28 May, 2025 3 commits
- [CI] Fix setup of disaggregation with different tp (#6706) · d63e76f7
  shangmingc authored May 29, 2025
```
Signed-off-by: Shangming Cai <caishangming@linux.alibaba.com>
```
  d63e76f7
- [CI] Fix flaky pp single node test (#6689) · c25231c6
  shangmingc authored May 28, 2025
```
Signed-off-by: Shangming Cai <caishangming@linux.alibaba.com>
```
  c25231c6
- Update amd docker and nightly models. (#6687) · f4a8987f
  Sai Enduri authored May 28, 2025
  
  f4a8987f
27 May, 2025 4 commits
- feat: Add warnings for invalid tool_choice and UTs (#6582) · 41ba767f
  Chang Su authored May 27, 2025
  
  41ba767f
- Add batch test for draft extend (#6672) · f127355a
  Ke Bao authored May 28, 2025
  
  f127355a
- Support gathering expert distribution details (#6665) · 87068b5c
  fzyzcjy authored May 28, 2025
  
  87068b5c
- [CI] update verlengine ci to 4-gpu test (#6007) · 2103b806
  Junrong Lin authored May 28, 2025
  
  2103b806
26 May, 2025 7 commits
- Update nightly thresholds and dependencies. (#6635) · eb8f02dd
  Sai Enduri authored May 26, 2025
  
  eb8f02dd
- Supported precomputed feature for Kimi VL (#6599) · 0d503090
  Lifu Huang authored May 26, 2025
  
  0d503090
- qwen3moe support two batch overlap (#6598) · f9bab3d5
  Yi Zhang authored May 26, 2025
  
  f9bab3d5
- feat: Improve Mistral and Qwen25 function call parsing (#6597) · 16f69b1f
  Chang Su authored May 25, 2025
  
  16f69b1f
- refactor qwen moe code, use communicator to support tp+dp (#6581) · 65f09131
  Yi Zhang authored May 26, 2025
  
  65f09131
- chore: upgrade transformers 4.52.3 (#6575) · 7eb9d8e5
  Yineng Zhang authored May 25, 2025
```
Co-authored-by: Mick <mickjagger19@icloud.com>
```
  7eb9d8e5
- Improve performance of two batch overlap in some imbalanced cases (#6593) · a191a0e4
  fzyzcjy authored May 26, 2025
  
  a191a0e4
25 May, 2025 4 commits
- added support for tied weights in qwen pipeline parallelism (#6546) · 3f23d8cd
  Shenggui Li authored May 25, 2025
  
  3f23d8cd
- Support Phi-4 Multi-Modal (text + vision only) (#6494) · 022012aa
  Lifu Huang authored May 24, 2025
  
  022012aa
- Refactor vlm embedding routine to use precomputed feature (#6543) · 681fdc26
  Xinyuan Tong authored May 24, 2025
```
Signed-off-by: Xinyuan Tong <justinning0323@outlook.com>
```
  681fdc26
- Support overlapping two batches (#4068) · 0d477880
  fzyzcjy authored May 25, 2025
  
  0d477880
24 May, 2025 1 commit
- Fix GPU OOM (#6564) · 7a5e6ce1
  kk authored May 25, 2025
```
Co-authored-by: michael <michael.zhang@amd.com>
```
  7a5e6ce1