Skip to content

[build] Sync sglang-kt submodule (v0.7.0) #1

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

Open
github-actions wants to merge 1 commit into
base: main
Choose a base branch
from
Open

[build] Sync sglang-kt submodule (v0.7.0) #1

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion third_party/sglang
Open in desktop
Submodule sglang updated 88 files
+218 −0 examples/chat_template/glm5_next_multimodal.jinja
+1 −1 python/pyproject.toml
+462 −0 python/sglang/jit_kernel/csrc/nsa/kpool_topk_transform.cuh
+80 −0 python/sglang/jit_kernel/kpool_topk_transform.py
+284 −0 python/sglang/kernels/ops/layernorm/mhc.py
+3 −0 python/sglang/srt/configs/__init__.py
+629 −0 python/sglang/srt/configs/glm5_next.py
+26 −3 python/sglang/srt/configs/model_config.py
+1 −0 python/sglang/srt/entrypoints/openai/protocol.py
+25 −12 python/sglang/srt/entrypoints/openai/serving_chat.py
+6 −0 python/sglang/srt/function_call/base_format_detector.py
+58 −1 python/sglang/srt/function_call/function_call_parser.py
+85 −1 python/sglang/srt/function_call/glm47_moe_detector.py
+26 −0 python/sglang/srt/layers/attention/attention_registry.py
+17 −0 python/sglang/srt/layers/attention/hybrid_linear_attn_backend.py
+214 −0 python/sglang/srt/layers/attention/linear/glm5_next_kda_backend.py
+122 −0 python/sglang/srt/layers/attention/linear/kernels/glm5_next_kda.py
+536 −0 python/sglang/srt/layers/attention/linear/kernels/glm5_next_kda_ops.py
+353 −0 python/sglang/srt/layers/attention/nsa/glm5_next_indexer_logits.py
+386 −0 python/sglang/srt/layers/attention/nsa/glm5_next_indexer_triton.py
+150 −0 python/sglang/srt/layers/attention/nsa/glm5_next_scaled_fp8.py
+444 −0 python/sglang/srt/layers/attention/nsa/glm5_next_sparse_attention.py
+1,943 −0 python/sglang/srt/layers/attention/nsa/kpool_fp8_index.py
+629 −0 python/sglang/srt/layers/attention/nsa/kpool_plan.py
+1,154 −0 python/sglang/srt/layers/attention/nsa/nsa_indexer_kpool.py
+19 −2 python/sglang/srt/layers/attention/nsa/utils.py
+382 −61 python/sglang/srt/layers/attention/nsa_backend.py
+22 −0 python/sglang/srt/layers/attention/utils.py
+8 −3 python/sglang/srt/layers/communicator.py
+357 −0 python/sglang/srt/layers/communicator_mhc.py
+29 −1 python/sglang/srt/layers/moe/fused_moe_triton/fused_moe.py
+2 −0 python/sglang/srt/layers/moe/fused_moe_triton/layer.py
+159 −0 python/sglang/srt/layers/moe/glm5_next_swiglu.py
+1,723 −118 python/sglang/srt/layers/moe/kt_ep_wrapper.py
+4 −0 python/sglang/srt/layers/moe/moe_runner/base.py
+23 −1 python/sglang/srt/layers/moe/moe_runner/triton.py
+12 −1 python/sglang/srt/managers/forward_hooks_registry.py
+139 −0 python/sglang/srt/managers/glm5_next_kpool_coordinator.py
+23 −4 python/sglang/srt/managers/schedule_batch.py
+4 −0 python/sglang/srt/managers/scheduler.py
+18 −6 python/sglang/srt/managers/scheduler_output_processor_mixin.py
+28 −0 python/sglang/srt/managers/tokenizer_manager.py
+595 −0 python/sglang/srt/mem_cache/glm5_next_memory_pool.py
+40 −2 python/sglang/srt/mem_cache/memory_pool.py
+32 −14 python/sglang/srt/model_executor/cuda_graph_runner.py
+16 −0 python/sglang/srt/model_executor/forward_batch_info.py
+46 −0 python/sglang/srt/model_executor/model_runner.py
+146 −29 python/sglang/srt/model_executor/model_runner_kv_cache_mixin.py
+2 −0 python/sglang/srt/models/deepseek_v2.py
+1,474 −0 python/sglang/srt/models/glm5_next.py
+236 −0 python/sglang/srt/models/glm5_next_dsa.py
+261 −0 python/sglang/srt/models/glm5_next_moe.py
+32 −0 python/sglang/srt/models/glm5_next_norm.py
+12 −1 python/sglang/srt/models/glm_ocr.py
+798 −0 python/sglang/srt/multimodal/processors/glm5_next.py
+285 −6 python/sglang/srt/server_args.py
+49 −1 python/sglang/srt/utils/hf_transformers_utils.py
+636 −0 scripts/glm5_next_consumer_gpu_compile.py
+835 −0 scripts/glm5_next_session_c_benchmark.py
+1,941 −0 scripts/glm5_next_session_c_final_acceptance.py
+1,157 −0 scripts/glm5_next_session_c_graph_acceptance.py
+1,157 −0 scripts/glm5_next_session_c_oracle.py
+1,345 −0 scripts/glm5_next_session_c_qa_gsm_graph_acceptance.py
+124 −0 test/registered/function_call/test_glm5_next_tool_choice.py
+167 −0 test/registered/kernels/ops/layernorm/test_mhc_flat.py
+650 −0 test/registered/kernels/test_dsa_kpool_multi_pool.py
+460 −0 test/registered/unit/test_communicator_mhc_state.py
+549 −0 test/registered/unit/test_glm5_next_consumer_gpu_kernels.py
+278 −0 test/registered/unit/test_glm5_next_dsa.py
+227 −0 test/registered/unit/test_glm5_next_indexer_logits.py
+673 −0 test/registered/unit/test_glm5_next_kda.py
+256 −0 test/registered/unit/test_glm5_next_kda_fp32_o_proj.py
+498 −0 test/registered/unit/test_glm5_next_kpool_pool_lifecycle.py
+618 −0 test/registered/unit/test_glm5_next_loader_contract.py
+686 −0 test/registered/unit/test_glm5_next_moe.py
+1,102 −0 test/registered/unit/test_glm5_next_regression_guards.py
+156 −0 test/registered/unit/test_glm5_next_runtime_config_subprocess.py
+711 −0 test/registered/unit/test_glm5_next_server_args_boundary.py
+268 −0 test/registered/unit/test_glm5_next_session_c_benchmark.py
+649 −0 test/registered/unit/test_glm5_next_session_c_final_acceptance.py
+390 −0 test/registered/unit/test_glm5_next_session_c_graph_acceptance.py
+457 −0 test/registered/unit/test_glm5_next_session_c_oracle.py
+486 −0 test/registered/unit/test_glm5_next_session_c_qa_gsm_graph_acceptance.py
+221 −0 test/registered/unit/test_glm5_next_session_d_contract.py
+99 −0 test/registered/unit/test_glm5_next_session_d_processor_subprocess.py
+469 −0 test/registered/unit/test_glm5_next_sparse_attention.py
+1,293 −0 test/registered/unit/test_glm5_next_stage3.py
+1,327 −0 test/unit/test_glm5_next_fp8_layerwise_prefill.py