sglang

ascend

[Feature] npu support enable_torch_compile for torchair backend (#13410 )

2025-12-16 09:23:51 +08:00

configs

…

cpu

[CPU] Add 4D input support for ROPE in sgl-kernel (#9337 )

2025-12-16 17:27:39 +08:00

debug_utils

…

ep

…

external_models

…

hicache

…

layers/attention/mamba

…

models

[Qwen3-next] support mamba radix cache for overlap scheduler (#14792 )

2025-12-14 18:54:16 -08:00

nightly

…

openai_server

…

ops

…

quant

…

rl

…

rotary_embedding

…

xpu

…

double-sparsity-config-Llama-3.1-8B-Instruct.json

…

experiment_runner.py

…

kv_cache_scales_llama3_1_8b.json

…

kv_cache_scales_llama3_8b.json

…

kv_cache_scales_qwen2_1_5b.json

…

parse_results.py

…

run_suite.py

Split test_piecewise_cuda_graph.py to optimize CI resource usage (#15290 )

2025-12-17 21:20:21 -08:00

test_abort.py

…

test_bench_one_batch.py

…

test_bench_serving.py

…

test_bench_typebaseddispatcher.py

…

test_bnb.py

…

test_chunked_prefill.py

…

test_constrained_decoding_spec_reasoning.py

…

test_constrained_decoding.py

…

test_create_kvindices.py

…

test_cutedsl_moe.py

…

test_data_parallelism.py

…

test_deepseek_v3_basic.py

…

test_deepseek_v3_cutedsl_4gpu.py

…

test_deepseek_v3_fp4_4gpu.py

[Perf] Enable Flashinfer autotune by default (#14357 )

2025-12-16 23:01:39 -08:00

test_deepseek_v3_mtp.py

…

test_deepseek_v32_basic.py

…

test_deepseek_v32_mtp.py

…

test_deterministic.py

…

test_disaggregation_basic.py

…

test_disaggregation_different_tp.py

…

test_disaggregation_dp_attention.py

…

test_disaggregation_hybrid_attention.py

[Qwen3-next] Add PD disaggregation support for mamba with extra_buffer (#15180 )

2025-12-16 14:36:00 +08:00

test_disaggregation_pp.py

…

test_dp_attention.py

…

test_eagle_dp_attention.py

…

test_eagle_infer_beta_dp_attention.py

…

test_epd_disaggregation.py

fix: moving decorator to header (#15297 )

2025-12-16 17:05:45 -08:00

test_eval_accuracy_large.py

…

test_eval_fp8_accuracy.py

…

test_external_models.py

tiny unify environ usage (#15335 )

2025-12-17 23:31:43 +08:00

test_fa3.py

…

test_flash_attention_4.py

…

test_flashmla.py

Reserve more memory for DeepSeekOCR model and adjust server start timeout for DeepGEMM to reduce flakiness (#15277 )

2025-12-17 13:13:05 -08:00

test_fp8_blockwise_gemm.py

…

test_fp8_utils.py

…

test_fused_moe.py

…

test_gguf.py

…

test_gpt_oss_1gpu.py

…

test_gpt_oss_4gpu.py

…

test_gpt_oss_common.py

…

test_gptqmodel_dynamic.py

…

test_harmony_parser.py

…

test_hidden_states.py

…

test_hybrid_attn_backend.py

…

test_input_embeddings.py

…

test_io_struct.py

…

test_jinja_template_utils.py

…

test_llama31_fp4.py

…

test_load_weights_from_remote_instance.py

support non disturbing remote instance weight loader v2 (#14997 )

2025-12-16 14:39:56 -08:00

test_local_attn.py

…

test_mamba_unittest.py

[Qwen3-next] support mamba radix cache for overlap scheduler (#14792 )

2025-12-14 18:54:16 -08:00

test_metrics_utils.py

…

test_metrics.py

…

test_mistral_large3_basic.py

…

test_mla_deepseek_v3.py

Increase timeout for TestDeepseekV3MTP for potential DeepGEMM cold start (#15239 )

2025-12-16 00:26:10 -08:00

test_mla_flashinfer.py

…

test_mla_fp8.py

…

test_mla_int8_deepseek_v3.py

…

test_mla.py

…

test_model_hooks.py

…

test_modelopt_export.py

…

test_modelopt_loader.py

…

test_moe_eval_accuracy_large.py

…

test_multi_instance_release_memory_occupation.py

…

test_multi_tokenizer.py

…

test_ngram_speculative_decoding.py

…

test_no_chunked_prefill.py

…

test_no_overlap_scheduler.py

…

test_original_logprobs.py

…

test_page_size.py

…

test_patch_torch.py

…

test_penalty.py

…

test_piecewise_cuda_graph_1_gpu_a.py

Split test_piecewise_cuda_graph.py to optimize CI resource usage (#15290 )

2025-12-17 21:20:21 -08:00

test_piecewise_cuda_graph_1_gpu_b.py

Split test_piecewise_cuda_graph.py to optimize CI resource usage (#15290 )

2025-12-17 21:20:21 -08:00

test_piecewise_cuda_graph_2_gpu.py

Split test_piecewise_cuda_graph.py to optimize CI resource usage (#15290 )

2025-12-17 21:20:21 -08:00

test_pp_single_node.py

[PP] Add pp support for Qwen3-VL (#12333 )

2025-12-17 16:03:58 +08:00

test_priority_scheduling.py

…

test_profile_merger_http_api.py

…

test_profile_merger.py

…

test_profile_v2.py

…

test_pytorch_sampling_backend.py

…

test_quantization.py

Fix import warnings (#15144 )

2025-12-14 21:25:18 -08:00

test_radix_attention.py

Enhance runtime memory check in CI (#15192 )

2025-12-16 21:40:38 +08:00

test_radix_cache_unit.py

…

test_reasoning_parser.py

…

test_release_memory_occupation.py

…

test_request_queue_validation.py

…

test_retract_decode.py

Enhance runtime memory check in CI (#15192 )

2025-12-16 21:40:38 +08:00

test_rope_rocm.py

…

test_score_api.py

…

test_server_args.py

…

test_skip_tokenizer_init.py

…

test_speculative_registry.py

…

test_srt_endpoint.py

…

test_srt_engine.py

…

test_standalone_speculative_decoding.py

…

test_start_profile.py

…

test_swa_unittest.py

…

test_torch_compile_moe.py

…

test_torch_compile.py

…

test_torch_native_attention_backend.py

…

test_torchao.py

…

test_triton_attention_backend.py

…

test_triton_attention_kernels.py

…

test_triton_fused_moe.py

…

test_triton_moe_channel_fp8_kernel.py

…

test_triton_sliding_window.py

…

test_type_based_dispatcher.py

…

test_utils_update_weights.py

…

test_video_utils.py

…

test_vision_chunked_prefill.py

Increase timeout for TestDeepseekV3MTP for potential DeepGEMM cold start (#15239 )

2025-12-16 00:26:10 -08:00

test_vision_openai_server_a.py

Reserve more memory for DeepSeekOCR model and adjust server start timeout for DeepGEMM to reduce flakiness (#15277 )

2025-12-17 13:13:05 -08:00

test_vision_openai_server_common.py

Reserve more memory for DeepSeekOCR model and adjust server start timeout for DeepGEMM to reduce flakiness (#15277 )

2025-12-17 13:13:05 -08:00

test_vlm_input_format.py

…

test_wave_attention_kernels.py

…