Skip to content

Issues: sgl-project/sglang

Development Roadmap (2025 H1)
#4042 opened Mar 4, 2025 by zhyncs
Open 15
[Roadmap] Prefill and Decoding Disaggregation
#4655 opened Mar 21, 2025 by ByronHsu
Open 18
[Roadmap] EP Enhancement
#4734 opened Mar 24, 2025 by ch-wan
Open 6
Beta
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

498 Open 1,276 Closed
498 Open 1,276 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Issues list

[Usage] How to run multiple model instances on a single GPU
#5507 opened Apr 17, 2025 by spitzblattr
[Bug] Incorrect passing of ForwardBatch parameter in TpModelWorker.forward_batch_generation
#5506 opened Apr 17, 2025 by u4lr451
2 of 5 tasks
@hebiao064 @Fridge003
Large Discrepancy in Speedup Between SGLang + Eagle and Eagle Repo Code
#5502 opened Apr 17, 2025 by motigrez
[BUG] some problems with HiRadixCache
#5499 opened Apr 17, 2025 by a4zhangfei
[Bug] Qwen-gme embedding model: cannot get fused embedding from text+image, and image input format may be incorrect
#5498 opened Apr 17, 2025 by wwqq
5 tasks done
[Bug] cannot import name 'ck_moe_2stages_win4' from 'aiter.fused_moe_bf16_asm' (/root/miniconda3/envs/xinf/lib/python3.10/site-packages/aiter/fused_moe_bf16_asm.py)
#5494 opened Apr 17, 2025 by githust66
5 tasks
1
[Feature] Is it currently supported to use two machines to form a single P or a single D? For example, deploying DeepSeek-R1 requires two H100s to form one instance.
#5492 opened Apr 17, 2025 by ZhenshengWu
2 tasks done
[Bug] sglang crush during run with eagle decode
#5491 opened Apr 17, 2025 by liuteng
1 of 5 tasks
[performance] 1P1D deployment of DeepSeek-R1 on two H200 machines does not meet performance expectations.
#5490 opened Apr 17, 2025 by ZhenshengWu
5 tasks done
1
[Bug] Potentially create too much process when using verl-sglang init_model
#5483 opened Apr 17, 2025 by SwordFaith
5 tasks done
2
[Feature] Optimize the Dockerfile for the project
#5474 opened Apr 16, 2025 by whybeyoung
2 tasks
2
[RFC][Feature][Model] Add templated fallback HF transformers model backend in SRT
#5471 opened Apr 16, 2025 by XuehaiPan
1 of 6 tasks
@XuehaiPan
3
[Bug] SGLang Server Freezes During High Traffic Periods with 16-GPU DeepSeek v3 Setup
#5465 opened Apr 16, 2025 by moqimoqidea
5 tasks done
[Bug] 模型一直停留在Prefill batch阶段，无法进行decode 出现卡死情况，最终导致计算节点Watchdog Timeout
#5458 opened Apr 16, 2025 by WALLE-AI
5 tasks
3
[Bug] rank 0 could finish the model loading, but there are other ranks that didn't finish loading. It is likely due to unexpected failures (eg,OOM) or a slow node.
#5457 opened Apr 16, 2025 by Alex-9827
5 tasks done
[Bug] incorrect output from same-prefix requests for Qwen2.5VL
#5455 opened Apr 16, 2025 by KivenChen
5 tasks done
3
What is the relationship between ModelRunner and the model(deepseek.py,llama.py..etc)?
#5453 opened Apr 16, 2025 by wangzhen2271
[Bug] sglang.bench_serving IndexError when add extra body stream_options.include_usage
#5451 opened Apr 16, 2025 by morty-zxb
5 tasks done
[Bug] PD disaggregation, KV transfer slow down under high concurrency
#5450 opened Apr 16, 2025 by MtFitzRoy
5 tasks done
12
[Bug] run eagle3 failed
#5448 opened Apr 16, 2025 by riou-chen
2
[Feature] disable-req-waiting
#5446 opened Apr 16, 2025 by voidxb
2 tasks
2
[Bug] v0.4.5 NameError: name 'VLLM_AVAILABLE' is not defined from compressed_tensors.py
#5443 opened Apr 16, 2025 by kratorado
5 tasks done
1
[Bug] ValueError: Model architectures ['Glm4ForCausalLM'] are not supported for now.
#5441 opened Apr 16, 2025 by chunxingque
5 tasks done
2
[Feature] optimize SegmentPackBits high priority speculative-decoding
#5437 opened Apr 15, 2025 by zhyncs
2 tasks
1
Help Needed: Slow Inference on 2xH100x8 Setup with DeepSeek-R1
#5429 opened Apr 15, 2025 by seungduk-yanolja
1
ProTip! Adding no:label will show everything without a label.