sglang

test_deepseek_mla_dispatch.py

[Refactor] Separate ROCm-specific DeepSeek MHA and MLA forward paths (#31531 )

2026-08-08 13:39:36 -07:00

test_deepseek_v4_amd_fused_mhc.py

[Config] Round 5.1: the published-side readers ask the bags, and a platform fact gets one address (#37086 )

2026-08-30 02:18:33 -07:00

test_deepseek_v4_amd_wo_a_bf16.py

[AMD] DeepSeek-V4: route decode wo_a bf16 batched matmul to aiter batched_gemm_bf16 (#33313 )

2026-08-19 03:04:48 -07:00

test_deepseek_v4_mxfp4_shared_expert_requant.py

[Deepseek-V4] Enable shared-experts fusion on the flashinfer_mxfp4 (trtllm-gen) MoE path (#35505 )

2026-08-25 14:55:25 -07:00

test_deepseek_v4_rope_policy.py

[Fix] Restore layer-level DSV4 RoPE policy (#34788 )

2026-08-13 21:26:45 -07:00

test_deepseek_v4_shared_expert_fusion.py

Fix DSV4 shared-fusion CPU unit test after the EP guard landed (#36424 )

2026-08-25 19:52:05 -07:00

test_draft_entry_hook_parity.py

config: each runner carries its own linear-attn kernel choice

2026-08-15 00:36:04 -07:00

test_flux2_fp8_norm_quant_gate.py

[Diffusion] Fuse FLUX.2 ModelOpt FP8 producers and QKV packing (#37162 )

2026-09-01 16:14:29 +08:00

test_fusion_gate_coverage.py

moe: the shared-experts-fusion decision is a per-runner value the loader installs (#33889 )

2026-08-07 22:42:58 -07:00

test_glm_moe_gate_fp32.py

Fix stale GLM MoE routing after runtime weight updates (#35883 )

2026-08-30 14:13:34 -07:00

test_hunyuan_v3_nextn_weight_loading.py

[Fix] Load HunyuanV3 NextN final_layernorm into the draft head's output norm (#30331 )

2026-07-13 20:37:36 +08:00

test_kimi_k3_bfa_overlap.py

[Fix] Support Kimi-K3 ModelOpt mixed NVFP4/FP8 checkpoint (#35077 )

2026-08-19 08:13:45 -07:00

test_kimi_k3_vision.py

[CI][RFC] Replace black-jupyter with ruff-format (#37210 )

2026-09-02 19:46:08 -07:00

test_kimi_k25_mm_projection.py

[Kimi] Support kimi-k3 (#32541 )

2026-08-04 13:22:49 -07:00

test_kimi_k25.py

config: a parallel size has one spelling; a patched scope declares its own (#36621 )

2026-08-27 12:56:42 -07:00

test_kimi_vl_moonvit.py

fix: fix Kimi-VL encoder parallelism (#30869 )

2026-07-14 08:44:06 +08:00

test_kimi_vl.py

[CI][RFC] Replace black-jupyter with ruff-format (#37210 )

2026-09-02 19:46:08 -07:00

test_llava_processor_pool.py

[vlm] fix: recover multimodal decode and processor failures (#36983 )

2026-08-30 20:50:19 +08:00

test_llava.py

[vlm] fix: recover multimodal decode and processor failures (#36983 )

2026-08-30 20:50:19 +08:00

test_locate_anything.py

chore: cleanup garbage code (#29770 )

2026-07-02 16:14:01 +08:00

test_longcat_flash_router_hpc_gemm.py

Optimize LongCat-Flash router GEMM with the HPC-Ops bf16xfp32 kernel (#30247 )

2026-07-21 20:05:17 +08:00

test_mellum.py

[Model] Add support for JetBrains' Mellum v2 code generation model (#27375 )

2026-07-13 22:54:38 -07:00

test_moss_vl_processor.py

[vlm] fix: recover multimodal decode and processor failures (#36983 )

2026-08-30 20:50:19 +08:00

test_nemotron_h_weight_loading.py

[NemotronH] Load shared embed_tokens/lm_head in MTP draft weights (#30456 )

2026-07-16 17:20:38 +00:00

test_paddleocr_vl_serving_defaults.py

[VLM] feat: size the multimodal preprocessing pool by where preprocessing runs (#35349 )

2026-08-27 16:32:58 +08:00

test_paddleocr_vl_vision.py

[perf] overlap page preprocessing, pack the vit, enable prefill CUDA graph for paddle-ocr (#35318 )

2026-08-19 08:20:55 +08:00

test_qwen3_5_modelopt_fp4.py

Qwen3.5-MoE: support modelopt_fp4 checkpoints that quantize attention (+ load baked FP8 KV scales) (#31220 )

2026-07-30 14:30:26 -07:00

test_qwen3_5_packed_weight_loader.py

[CI] Rename basic CI stage-a/b/c -> base-a/b/c for symmetry with extra CI (#25420 )

2026-05-15 18:26:55 -07:00

test_qwen3_5_pipeline_parallel.py

[Bugfix] Load Qwen3.5 MTP embedding under PP (#37471 )

2026-09-02 15:19:40 -07:00

test_qwen3_embedding_registration.py

model: serve bare Qwen3Model backbone natively as an embedding model (#32457 )

2026-07-27 15:49:58 +08:00

test_qwen3_vl_feature_materialization.py

[CI][RFC] Replace black-jupyter with ruff-format (#37210 )

2026-09-02 19:46:08 -07:00

test_qwen_image_fp8_norm_quant.py

[Diffusion] Fuse Qwen-Image FP8 norm and activation quantization (#37156 )

2026-08-31 18:19:42 +08:00

test_shared_experts_fusion_gates.py

config: a parallel leaf with no live counterpart is read bare (#36620 )

2026-08-27 12:56:11 -07:00

test_vit_pos_embed_interpolate.py

[CI] Move CPU-only unit tests to the CPU suite and trim dead 5090 registrations (#33654 )

2026-08-05 01:43:29 -07:00

test_xllm.py

[Model] Add native IFM K2 Horizon serving support (#37654 )

2026-09-03 16:39:43 +08:00

test_zaya_cca.py

config: a parallel size has one spelling; a patched scope declares its own (#36621 )

2026-08-27 12:56:42 -07:00

test_zaya_mod_tp.py

Support for Zyphra zaya1 model (#26347 )

2026-06-10 02:44:47 -07:00