Search results

unit-test-deepep-8-gpu: test_deepep_large.py failing with IBGDA/cudaHostRegister errors

Status: Open.

#17175 In sgl-project/sglang;

· alisonshao opened on Jan 15, 2026

[Bug] DeepSeek-V3.2-NVFP4 shows garbled output after EPLB

Status: Open.

#17162 In sgl-project/sglang;

· wyzzbond opened on Jan 15, 2026

[Bug] Qwen3 Eagle3 on NPU fails with split_qkv_rmsnorm_rope argument error

Status: Open.

#17153 In sgl-project/sglang;

· zzzzzzzxh opened on Jan 15, 2026

test_tensor_dump_forward_hook.py uses pytest-style function without TestCase class

Status: Open.

#17145 In sgl-project/sglang;

· alisonshao opened on Jan 15, 2026

[Bug] Qwen3-Coder-480B-A35B-Instruct-FP8 with flashinfer-trtllm kernel results are not correct.

Status: Open.

#17134 In sgl-project/sglang;

· jiapingW opened on Jan 15, 2026

SGLang Nvidia Collaboration Roadmap (2026 Q1)

Status: Open.

#17130 In sgl-project/sglang;

· Fridge003 opened on Jan 15, 2026

[RFC] sgl-kernel Release Strategy Under Current Storage Quota Constraints

dependenciesPull requests that update a dependency filePull requests that update a dependency file

high priority

Status: Open.

#17110 In sgl-project/sglang;

· zhyncs opened on Jan 15, 2026

[Feature] How to use Claude Code for sglang, is there some specific prompt?

Status: Open.

#17109 In sgl-project/sglang;

· yurikokatayama1539 opened on Jan 15, 2026

[Bug] DeepSeek 3.1T update weights error

Status: Open.

#17107 In sgl-project/sglang;

· haozhihuiYmh150 opened on Jan 15, 2026

[Feature] Expose block_size and num_gpu_blocks metrics to enable downstream autotuning

Status: Open.

#17104 In sgl-project/sglang;

· BenjaminBraunDev opened on Jan 15, 2026

[Bug] Reduce SGLang diffusion generate scheduler overhead

Status: Open.

#17102 In sgl-project/sglang;

· BBuf opened on Jan 15, 2026

[Bug] Dynamic LoRA load error

Status: Open.

#17096 In sgl-project/sglang;

· rulongchen opened on Jan 15, 2026