sglang 0.4.9.post4__py3-none-any.whl → 0.4.9.post6__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- sglang/lang/chat_template.py +21 -0
- sglang/srt/configs/internvl.py +3 -0
- sglang/srt/configs/model_config.py +7 -0
- sglang/srt/constrained/base_grammar_backend.py +10 -2
- sglang/srt/constrained/xgrammar_backend.py +7 -5
- sglang/srt/conversation.py +16 -1
- sglang/srt/debug_utils/__init__.py +0 -0
- sglang/srt/debug_utils/dump_comparator.py +131 -0
- sglang/srt/debug_utils/dumper.py +108 -0
- sglang/srt/debug_utils/text_comparator.py +172 -0
- sglang/srt/disaggregation/decode_schedule_batch_mixin.py +13 -1
- sglang/srt/disaggregation/mooncake/conn.py +16 -0
- sglang/srt/disaggregation/prefill.py +13 -1
- sglang/srt/entrypoints/engine.py +4 -2
- sglang/srt/entrypoints/http_server.py +13 -1
- sglang/srt/entrypoints/openai/protocol.py +3 -1
- sglang/srt/entrypoints/openai/serving_base.py +5 -2
- sglang/srt/entrypoints/openai/serving_chat.py +132 -79
- sglang/srt/function_call/ebnf_composer.py +10 -3
- sglang/srt/function_call/function_call_parser.py +2 -0
- sglang/srt/function_call/glm4_moe_detector.py +164 -0
- sglang/srt/function_call/qwen3_coder_detector.py +1 -0
- sglang/srt/layers/attention/hybrid_attn_backend.py +100 -0
- sglang/srt/layers/attention/vision.py +56 -8
- sglang/srt/layers/layernorm.py +26 -1
- sglang/srt/layers/logits_processor.py +14 -3
- sglang/srt/layers/moe/ep_moe/layer.py +323 -242
- sglang/srt/layers/moe/ep_moe/token_dispatcher.py +83 -118
- sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=160,N=320,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=160,N=192,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=160,N=384,device_name=NVIDIA_H200,dtype=fp8_w8a8.json +146 -0
- sglang/srt/layers/moe/fused_moe_triton/layer.py +38 -48
- sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py +11 -8
- sglang/srt/layers/moe/token_dispatcher/__init__.py +0 -0
- sglang/srt/layers/moe/token_dispatcher/base_dispatcher.py +48 -0
- sglang/srt/layers/moe/token_dispatcher/standard.py +19 -0
- sglang/srt/layers/moe/topk.py +90 -24
- sglang/srt/layers/multimodal.py +11 -8
- sglang/srt/layers/quantization/fp8.py +25 -247
- sglang/srt/layers/quantization/fp8_kernel.py +78 -48
- sglang/srt/layers/quantization/modelopt_quant.py +27 -10
- sglang/srt/layers/quantization/unquant.py +24 -76
- sglang/srt/layers/quantization/w4afp8.py +68 -17
- sglang/srt/lora/lora_registry.py +93 -29
- sglang/srt/managers/cache_controller.py +9 -7
- sglang/srt/managers/data_parallel_controller.py +4 -0
- sglang/srt/managers/io_struct.py +12 -0
- sglang/srt/managers/mm_utils.py +154 -35
- sglang/srt/managers/multimodal_processor.py +3 -14
- sglang/srt/managers/schedule_batch.py +14 -8
- sglang/srt/managers/scheduler.py +64 -1
- sglang/srt/managers/scheduler_input_blocker.py +106 -0
- sglang/srt/managers/tokenizer_manager.py +80 -15
- sglang/srt/managers/tp_worker.py +8 -0
- sglang/srt/mem_cache/hiradix_cache.py +5 -2
- sglang/srt/model_executor/model_runner.py +83 -27
- sglang/srt/models/deepseek_v2.py +75 -84
- sglang/srt/models/glm4_moe.py +1035 -0
- sglang/srt/models/glm4_moe_nextn.py +167 -0
- sglang/srt/models/interns1.py +328 -0
- sglang/srt/models/internvl.py +143 -47
- sglang/srt/models/llava.py +9 -5
- sglang/srt/models/minicpmo.py +4 -1
- sglang/srt/models/qwen2_moe.py +2 -2
- sglang/srt/models/qwen3_moe.py +17 -71
- sglang/srt/multimodal/processors/base_processor.py +20 -6
- sglang/srt/multimodal/processors/clip.py +2 -2
- sglang/srt/multimodal/processors/deepseek_vl_v2.py +2 -2
- sglang/srt/multimodal/processors/gemma3.py +2 -2
- sglang/srt/multimodal/processors/gemma3n.py +2 -2
- sglang/srt/multimodal/processors/internvl.py +21 -8
- sglang/srt/multimodal/processors/janus_pro.py +2 -2
- sglang/srt/multimodal/processors/kimi_vl.py +2 -2
- sglang/srt/multimodal/processors/llava.py +4 -4
- sglang/srt/multimodal/processors/minicpm.py +2 -3
- sglang/srt/multimodal/processors/mlama.py +2 -2
- sglang/srt/multimodal/processors/mllama4.py +18 -111
- sglang/srt/multimodal/processors/phi4mm.py +2 -2
- sglang/srt/multimodal/processors/pixtral.py +2 -2
- sglang/srt/multimodal/processors/qwen_audio.py +2 -2
- sglang/srt/multimodal/processors/qwen_vl.py +2 -2
- sglang/srt/multimodal/processors/vila.py +3 -1
- sglang/srt/poll_based_barrier.py +31 -0
- sglang/srt/reasoning_parser.py +2 -1
- sglang/srt/server_args.py +65 -6
- sglang/srt/two_batch_overlap.py +8 -3
- sglang/srt/utils.py +96 -1
- sglang/srt/weight_sync/utils.py +119 -0
- sglang/test/runners.py +4 -0
- sglang/test/test_utils.py +118 -5
- sglang/utils.py +19 -0
- sglang/version.py +1 -1
- {sglang-0.4.9.post4.dist-info → sglang-0.4.9.post6.dist-info}/METADATA +5 -4
- {sglang-0.4.9.post4.dist-info → sglang-0.4.9.post6.dist-info}/RECORD +97 -80
- sglang/srt/debug_utils.py +0 -74
- {sglang-0.4.9.post4.dist-info → sglang-0.4.9.post6.dist-info}/WHEEL +0 -0
- {sglang-0.4.9.post4.dist-info → sglang-0.4.9.post6.dist-info}/licenses/LICENSE +0 -0
- {sglang-0.4.9.post4.dist-info → sglang-0.4.9.post6.dist-info}/top_level.txt +0 -0
@@ -9,11 +9,11 @@ sglang/compile_deep_gemm.py,sha256=H118s76CKdpZr-cDeFfBCePe7--c_teEBNVIzchYVSo,6
|
|
9
9
|
sglang/global_config.py,sha256=xzLdk8W53fneFblNh8iIjGF9C3-7mnzR1-LleD9Btxg,1495
|
10
10
|
sglang/launch_server.py,sha256=mDXfwha8LHpWQJekcCosR98QhCQsbmilsBlI5jAIgg0,420
|
11
11
|
sglang/profiler.py,sha256=tEHzHerXC-ymk4OrkoUcMbgcGHmb8VESthsNSP2Yx9w,4417
|
12
|
-
sglang/utils.py,sha256=
|
13
|
-
sglang/version.py,sha256=
|
12
|
+
sglang/utils.py,sha256=DSMD5Ch-gmkIO6Dh4JAmBMtRAu4JZ7uYO0NUMprGVO4,16770
|
13
|
+
sglang/version.py,sha256=sV0w9abphYZ3w1RIYSGHC1mPd8_FLDhrgT8ogSNOvck,28
|
14
14
|
sglang/eval/llama3_eval.py,sha256=gWSboDchIGybIce88bJlrCG0yiLZ513mw4gcutJlzGM,10017
|
15
15
|
sglang/eval/loogle_eval.py,sha256=-CC2s2kh5qUoDrHRkQVkC_jNvBgNojXbf456ny5s78s,4557
|
16
|
-
sglang/lang/chat_template.py,sha256=
|
16
|
+
sglang/lang/chat_template.py,sha256=ZNcuDvDZE9781tQqfKpptck_zJVK0aggq7IxrNHltao,21934
|
17
17
|
sglang/lang/choices.py,sha256=-W1DVw9N9ZliVpvmWrzIXG4cswAah8eMQrHWzkS3D8o,6234
|
18
18
|
sglang/lang/compiler.py,sha256=MAuzoOOpb98njJ7Io2SDmFkhTroDYiq0te0ZpfHkMY4,7597
|
19
19
|
sglang/lang/interpreter.py,sha256=JneG-Yw-XyOTBZCIoxTiQluQjAlSz1YswnV0l6pqVPo,34811
|
@@ -30,20 +30,20 @@ sglang/srt/_custom_ops.py,sha256=9GPNUyqnybgvUnzdQ8n5_NgluHmQNw0Gptos0iLPfrY,534
|
|
30
30
|
sglang/srt/aio_rwlock.py,sha256=6LYtOdeTUY3hkfa1dmYkgsaF2ttrwIF3hUWz2AZ2fqw,2970
|
31
31
|
sglang/srt/code_completion_parser.py,sha256=KFa95OU0TeVjJkOqIgS4xV3kaJ5dFWOmAAgISyc1oEc,3803
|
32
32
|
sglang/srt/constants.py,sha256=0i-tEwG2BSYNDy96MxnGHV5HnBELkYcnsVGsE-R18o0,93
|
33
|
-
sglang/srt/conversation.py,sha256=
|
33
|
+
sglang/srt/conversation.py,sha256=ocpOVHIKV2C0Unp3ALwtW71HZo3tgaCzadXp62Omqdw,41604
|
34
34
|
sglang/srt/custom_op.py,sha256=7OPQAZe3PqDXk-xM0Wxl7pm9cazilNJfkNiYIWztWgI,3198
|
35
|
-
sglang/srt/debug_utils.py,sha256=slaFOY4BYDBFatkfu8FZlzai-u4LFS-5GUzdr-t50zE,2241
|
36
35
|
sglang/srt/hf_transformers_utils.py,sha256=UiGjp70ZpTaZDJ99Roc-2DTkxV3gR-2oJjjaNY9Ekh4,12391
|
37
36
|
sglang/srt/jinja_template_utils.py,sha256=0c6nQjkGukmyXsqQ9r85R5QK9CafnjqXyrXF4K3aFhE,7177
|
38
37
|
sglang/srt/model_parallel.py,sha256=eLXZhvJ4wG6dh0FontNCIdVZvHYdWgaeY-5cu7TD9tE,6078
|
39
38
|
sglang/srt/operations.py,sha256=ddQ8KO63L73OciaR8MZ9h2h83gKVY4-WuWgeEGowPJA,5346
|
40
39
|
sglang/srt/operations_strategy.py,sha256=Pwd2sKeRtKh9WJXgzlNr2tU9y6YMcI3MDLbatHqlMws,7145
|
41
40
|
sglang/srt/patch_torch.py,sha256=OUPCGQSQz3MVZB1zZ_Eq8lXiw0uIKJ_HWjqQolI8FsM,3088
|
42
|
-
sglang/srt/
|
43
|
-
sglang/srt/
|
41
|
+
sglang/srt/poll_based_barrier.py,sha256=9vjn32ROlSueEk7EuZjWKUrrL7XVR9KJsSVliGTYB_4,937
|
42
|
+
sglang/srt/reasoning_parser.py,sha256=6lkrkR8Q25EU3VAE0jjSq__361e3CkX0gJIieSfjsZk,9247
|
43
|
+
sglang/srt/server_args.py,sha256=-FRh5ZtXHAN2PVthw_LcH0KOgzlzeohFU24ZkeZBVm0,86265
|
44
44
|
sglang/srt/torch_memory_saver_adapter.py,sha256=K_eTx0UU84MHSTXI3iqYLdHV4IWtJMJ2FKdGFJR8v1E,2417
|
45
|
-
sglang/srt/two_batch_overlap.py,sha256=
|
46
|
-
sglang/srt/utils.py,sha256=
|
45
|
+
sglang/srt/two_batch_overlap.py,sha256=AgaPeAXJQjt4RVhIVKR2YMEZRL8NqjN93NpgDE_fXfQ,28892
|
46
|
+
sglang/srt/utils.py,sha256=6YkLoFlKyLpCNlk6zzaM5ys5_dMRTMOWNoLH09oEsOs,93321
|
47
47
|
sglang/srt/warmup.py,sha256=zldxhMlXpclRAJXmfBjJNUJd1eDizVdysibBvQyTVuA,1782
|
48
48
|
sglang/srt/configs/__init__.py,sha256=8EcVRP95epZ49DxBa6LgKWt7eO3Qe7Hrr3V1c6HkMnY,553
|
49
49
|
sglang/srt/configs/chatglm.py,sha256=j-b0YkdYUmQm2y1kNmMJtKeACxWKmBbvNNkDWbs6kbI,2907
|
@@ -51,12 +51,12 @@ sglang/srt/configs/dbrx.py,sha256=tdhIkXAQl1yr0MxqFmsDG1E0e2puRTTKm6UTyANBLac,11
|
|
51
51
|
sglang/srt/configs/deepseekvl2.py,sha256=sYMmoy5_LlDXht7MTi4QGYOaiE8iMGUO44bnuOu4i8I,23683
|
52
52
|
sglang/srt/configs/device_config.py,sha256=kdwFrk5myAURxdp4rSr8ANpBpSJfuBDoT-kuCyuscRs,442
|
53
53
|
sglang/srt/configs/exaone.py,sha256=Duxd4yQoKy8GWEzZD_kCY_OzmN_67CTJL_Kgn0eXk3g,10731
|
54
|
-
sglang/srt/configs/internvl.py,sha256=
|
54
|
+
sglang/srt/configs/internvl.py,sha256=meLG4M2s27bPI-yTleRimPdAuOH1DjiVwUf2U6vXzno,27844
|
55
55
|
sglang/srt/configs/janus_pro.py,sha256=yI2c_aWBIKiTcK6RZxK6hq8CIxyYRH_NN8uaSNKsrKI,19156
|
56
56
|
sglang/srt/configs/kimi_vl.py,sha256=4W7VQI3pr888ZsFA2SqCQo4mI0seXTOrGQ-x3oTvWew,1358
|
57
57
|
sglang/srt/configs/kimi_vl_moonvit.py,sha256=hx2Rt4JSFbvy2HUTeLjBpge87m8M6ITAhqsgdNf_Jd4,1163
|
58
58
|
sglang/srt/configs/load_config.py,sha256=qs-AxuplouBx2tsv9KGBOLZPbwzuVA4vbktbGP_cRp8,3309
|
59
|
-
sglang/srt/configs/model_config.py,sha256=
|
59
|
+
sglang/srt/configs/model_config.py,sha256=8lQ0tH3RORJAYSSosaS0jCPtuXkwcs6XUZsEQsZqgqM,28738
|
60
60
|
sglang/srt/configs/update_config.py,sha256=D-myV9Ombwp6wTF0RX9qUO6m5rxFiUcyse32FQADG28,4787
|
61
61
|
sglang/srt/configs/utils.py,sha256=3nHUfisMs_Ltuhv8OZTNCJp63YJKJVF43h1QZB1zqx8,670
|
62
62
|
sglang/srt/connector/__init__.py,sha256=czLX5JOxuMhH-T9eSJzoc1qv1B4z9chyffDRL5I6wo4,1247
|
@@ -67,19 +67,23 @@ sglang/srt/connector/utils.py,sha256=isTvxauz1_8no5MW6p8Bwj2H9mQkweaRO_RSwAPA7R8
|
|
67
67
|
sglang/srt/connector/serde/__init__.py,sha256=MvBJ7FBJtnou-AD-EdsCdAIDCcA8kWvUNuRViILVJ-0,718
|
68
68
|
sglang/srt/connector/serde/safe_serde.py,sha256=Iv_mFsDvk-cXuw0WTykBZ2yGxW0jb82DwVqC7r3RmAU,750
|
69
69
|
sglang/srt/connector/serde/serde.py,sha256=n59I2MXLa7WCyN_8pEd8L-scJk7lMhmEX-GOUIhF0ZA,1004
|
70
|
-
sglang/srt/constrained/base_grammar_backend.py,sha256=
|
70
|
+
sglang/srt/constrained/base_grammar_backend.py,sha256=ATWX1LKp3GjUDVkauEaSc3MnS8Bt6hOiHUgeYDZtgNo,7059
|
71
71
|
sglang/srt/constrained/llguidance_backend.py,sha256=Au8G8x4EB39w7k9-GFLP2hOPETyeAPw53RRZYdLimpA,6191
|
72
72
|
sglang/srt/constrained/outlines_backend.py,sha256=4Er45scnqHuJQ14kggsVX3RoHlaUD9t0ALIoWDQEmFg,6913
|
73
73
|
sglang/srt/constrained/outlines_jump_forward.py,sha256=Gyubp-FVetxd6wP4FA_kD6cCXIRfr8k_ZDviJyte048,6824
|
74
74
|
sglang/srt/constrained/reasoner_grammar_backend.py,sha256=YFxWuOTTo4e6cGhxnaBwuwli6f8FTUJtszib8dBq_8I,3207
|
75
|
-
sglang/srt/constrained/xgrammar_backend.py,sha256=
|
75
|
+
sglang/srt/constrained/xgrammar_backend.py,sha256=y_pmPUHEkqLM1BCHcakRYC99j2_hzrsivd5uJIKeuJE,8149
|
76
76
|
sglang/srt/constrained/triton_ops/bitmask_ops.py,sha256=WjTen9iuuFWLzkE1mAHQZB9_7aIy5QH8Wjf-lB-Fams,4614
|
77
|
+
sglang/srt/debug_utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
78
|
+
sglang/srt/debug_utils/dump_comparator.py,sha256=8ZPwP0zojfRHvPiSKgpESYFcC3tR-sBjIej5SotOmpg,3947
|
79
|
+
sglang/srt/debug_utils/dumper.py,sha256=eyrg5lmwyYF88kn8b06kpu1__aGnWb6CLM8fxywtooI,3158
|
80
|
+
sglang/srt/debug_utils/text_comparator.py,sha256=pg58q3fzsnd3rJKRgZTdcOviOtAv8pSYZdwf-TLffMM,5522
|
77
81
|
sglang/srt/disaggregation/decode.py,sha256=P-0OyFjjDfthEVhlJugdscWST_uoOVmHZlSlx9NYLmE,34242
|
78
|
-
sglang/srt/disaggregation/decode_schedule_batch_mixin.py,sha256=
|
82
|
+
sglang/srt/disaggregation/decode_schedule_batch_mixin.py,sha256=sUXmqiJGwuYoJvZjRCgTMVEN7bLefnEUlCTJpVA6ecI,6416
|
79
83
|
sglang/srt/disaggregation/kv_events.py,sha256=yFbtwOeblcCmOXTjg00TopxnyjkmCBQIVz46KB8jetY,13555
|
80
84
|
sglang/srt/disaggregation/launch_lb.py,sha256=mcbAztN4gnHevw_T5R2_nWsymsDEY9vHkm7OJ1vr6cc,4211
|
81
85
|
sglang/srt/disaggregation/mini_lb.py,sha256=bpnz6d90nv7vAoiPxmxFKlM1Gjct0R4IPkF9SwwBk9I,14504
|
82
|
-
sglang/srt/disaggregation/prefill.py,sha256=
|
86
|
+
sglang/srt/disaggregation/prefill.py,sha256=u8eKfWPuXiHDqFCWmhqlYhco3wy7nt5DjSmekwJTkhQ,24286
|
83
87
|
sglang/srt/disaggregation/utils.py,sha256=ql5p9yjBCRcR0YMgcgRK6tH0kPGvLNTUgeXBr_BUcoI,12198
|
84
88
|
sglang/srt/disaggregation/ascend/__init__.py,sha256=-lxnica6fZYmYNpKwmE8yESUpQ5Sxf2DiZoBHzboKc8,146
|
85
89
|
sglang/srt/disaggregation/ascend/conn.py,sha256=FOYrGtEhj58iTjhYT1eP96DpyxBLFFR87icqYCzWmCY,1215
|
@@ -92,7 +96,7 @@ sglang/srt/disaggregation/common/utils.py,sha256=SxRhAWisNK8seGhb5BXBJ5u53DF7yeK
|
|
92
96
|
sglang/srt/disaggregation/fake/__init__.py,sha256=jJGWdXwaQiGIoR6atKqkQfkJmVyQ09l55VUN2WjwaeY,77
|
93
97
|
sglang/srt/disaggregation/fake/conn.py,sha256=oD1DArn1yDFZCu-X6p93uSLlAXEkt9lYxERICMznxGw,2286
|
94
98
|
sglang/srt/disaggregation/mooncake/__init__.py,sha256=0TgqkAdQI1YynbHY6c0QISvVoOSk-0SwCIq5rjPSmgE,156
|
95
|
-
sglang/srt/disaggregation/mooncake/conn.py,sha256=
|
99
|
+
sglang/srt/disaggregation/mooncake/conn.py,sha256=4TnmzcCDDm15j3XFFw_n_awq958-ljVahy8nC3CuFdc,61842
|
96
100
|
sglang/srt/disaggregation/mooncake/transfer_engine.py,sha256=-ulLhz6DBDsmEjwNjjH5Pj8ngZKISgwBc_zL3ViGDN0,4761
|
97
101
|
sglang/srt/disaggregation/nixl/__init__.py,sha256=qODVPIGWUXKXq4zsRIcMYoAoAeg6nBIN9vdQOlVMANE,136
|
98
102
|
sglang/srt/disaggregation/nixl/conn.py,sha256=eSof87fG21Dd4COszfnbeXIxne3TWvw0mSvCOkjsBZc,20323
|
@@ -112,13 +116,13 @@ sglang/srt/distributed/device_communicators/quick_all_reduce.py,sha256=4j1_E4azo
|
|
112
116
|
sglang/srt/distributed/device_communicators/shm_broadcast.py,sha256=IrSrnpZnii0EJTS1CYRwEwE7gyHxJBVgI2QuJS3AKW0,20906
|
113
117
|
sglang/srt/distributed/device_communicators/xpu_communicator.py,sha256=ajW6132BvA6jkeipEIgN27TFycI0U06Ih2Z8WNjlA4s,1593
|
114
118
|
sglang/srt/entrypoints/EngineBase.py,sha256=yKN76witT2jz1zhmLHmPNLGMpK2UiOTaKQ2KPD8l99U,2594
|
115
|
-
sglang/srt/entrypoints/engine.py,sha256=
|
116
|
-
sglang/srt/entrypoints/http_server.py,sha256
|
119
|
+
sglang/srt/entrypoints/engine.py,sha256=dGazuzKYxQSfXjea1aeSqaZubvLER0EVZzD7CvpHq9A,31595
|
120
|
+
sglang/srt/entrypoints/http_server.py,sha256=aoXYyq5i4jfBppDrX5lECFFvsRlT90eW03gBXDTUpyQ,37647
|
117
121
|
sglang/srt/entrypoints/http_server_engine.py,sha256=_--j4U04OeJLlnnv1f0XmCd_Ry0z1FlhkrbePX8rYV0,4938
|
118
122
|
sglang/srt/entrypoints/openai/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
119
|
-
sglang/srt/entrypoints/openai/protocol.py,sha256=
|
120
|
-
sglang/srt/entrypoints/openai/serving_base.py,sha256=
|
121
|
-
sglang/srt/entrypoints/openai/serving_chat.py,sha256=
|
123
|
+
sglang/srt/entrypoints/openai/protocol.py,sha256=2XfURpoICsk4UZs4W9lm2A1yXNC-iwqSdZOuof0FXjo,19183
|
124
|
+
sglang/srt/entrypoints/openai/serving_base.py,sha256=epD6J7K72p6TP8o2dGYf-a76nXeF80vFhPRpWts6684,5445
|
125
|
+
sglang/srt/entrypoints/openai/serving_chat.py,sha256=MbLFLLFhvCTpJ_j7t03seXHHoFjlXCZtzVjSg_L7iz8,39841
|
122
126
|
sglang/srt/entrypoints/openai/serving_completions.py,sha256=emIutVmnJgOgGRuAqdB80-kVHUfQbE67n1VS-76o4QY,16690
|
123
127
|
sglang/srt/entrypoints/openai/serving_embedding.py,sha256=hqPt2ELpT1yQ8sBXo801aLJ3sExoPXs_K5ZQtC2vpAs,6274
|
124
128
|
sglang/srt/entrypoints/openai/serving_rerank.py,sha256=8n6y5kC6UhRpWrDtzH0eIp0dznW12FT60MLCR4bvD2U,3333
|
@@ -139,14 +143,15 @@ sglang/srt/eplb/eplb_simulator/reader.py,sha256=OlsWqsuXqVQY7PfvvrP-vsl2Ww2Kg4vD
|
|
139
143
|
sglang/srt/function_call/base_format_detector.py,sha256=FKBzsGFHfNl3Wvccsdvv9ftjbCZIyyjOLj2kHgKvX9g,15901
|
140
144
|
sglang/srt/function_call/core_types.py,sha256=wLUsW8mUA-i-ISz5QUDL7Ejq72W-K1HIhFa9Wm-_oA8,786
|
141
145
|
sglang/srt/function_call/deepseekv3_detector.py,sha256=WzJbwCe8DoOi8guX9pdVMd70CViVrwoaxi8PZwIC9pg,9712
|
142
|
-
sglang/srt/function_call/ebnf_composer.py,sha256=
|
143
|
-
sglang/srt/function_call/function_call_parser.py,sha256=
|
146
|
+
sglang/srt/function_call/ebnf_composer.py,sha256=HvLTgfNnBNKFKW-L36DKh07yRLbdTYjSfbTcLRUxZUc,14948
|
147
|
+
sglang/srt/function_call/function_call_parser.py,sha256=YTFGRA0rJf18ox2veLzmSBEm85QiZsPMNvNwiq8-iyo,8009
|
148
|
+
sglang/srt/function_call/glm4_moe_detector.py,sha256=TD67qS7T-9iR4-P5fhrUbqqYWEYVTb30dDNd7lGj4Rk,6934
|
144
149
|
sglang/srt/function_call/kimik2_detector.py,sha256=7unW_GwH7I6jrh2BZbw22mWRZzQ0OMTYDz_xVX5abPU,10116
|
145
150
|
sglang/srt/function_call/llama32_detector.py,sha256=trZja7IXc2IWUko5PRj7p4pRU6XCVjtxr-7qJNr0EWo,3647
|
146
151
|
sglang/srt/function_call/mistral_detector.py,sha256=Ts2HrJdu5r2lCCMVKPZfnCds9pt1K5H1EaOQrF_HD60,4686
|
147
152
|
sglang/srt/function_call/pythonic_detector.py,sha256=yAatItKtVDf9uhlB78wPWKP3fbefLJDNGGzFV5ey-YA,8756
|
148
153
|
sglang/srt/function_call/qwen25_detector.py,sha256=nv83_sipZMU6Vie3J4HeNiZVCSRU5ei7601j50779HA,5181
|
149
|
-
sglang/srt/function_call/qwen3_coder_detector.py,sha256=
|
154
|
+
sglang/srt/function_call/qwen3_coder_detector.py,sha256=3QAoK9W3YFPrx_j3xIsDpmQSnbX0BixMAlJ6BoNJo8o,5517
|
150
155
|
sglang/srt/function_call/utils.py,sha256=__ImDF2kNyoLWsYO5RYoryvy1mmgEjnjXlCvLv-uLCM,1695
|
151
156
|
sglang/srt/layers/activation.py,sha256=uhfhh10N5iLoLEBCvFPRjCfDxu5jXe1ehHJb3ISIjig,7840
|
152
157
|
sglang/srt/layers/amx_utils.py,sha256=1mENgHK2B8mgaD1oMtgbZ15Jmy_Uu1QueBmo09Ff2iA,2865
|
@@ -154,10 +159,10 @@ sglang/srt/layers/communicator.py,sha256=c5pJObNyP_7JsIWgLau_E-1ovtPCpAcge254fjo
|
|
154
159
|
sglang/srt/layers/dp_attention.py,sha256=tOre7il5Cppu930-tzM6d4AgjHuIJMCMlg9hJkuBcjA,11382
|
155
160
|
sglang/srt/layers/elementwise.py,sha256=MyQUflyKEfPZ-BggW1Kd4hB53RFD6FXGc2S5LXjx_do,16026
|
156
161
|
sglang/srt/layers/flashinfer_comm_fusion.py,sha256=fkTcAB7qYwSWi95qI3Rqq0JUyDpJdcYkd4TYkWO01HI,5891
|
157
|
-
sglang/srt/layers/layernorm.py,sha256=
|
162
|
+
sglang/srt/layers/layernorm.py,sha256=Qkx4Jzb3KWOt2okAVM6Is4FDClpUIJShcPGoElUSnEQ,9598
|
158
163
|
sglang/srt/layers/linear.py,sha256=NzjLqZrZNYcc3z2KtA9-n6pnSdBHHwU6NF_BbcAWp48,52471
|
159
|
-
sglang/srt/layers/logits_processor.py,sha256=
|
160
|
-
sglang/srt/layers/multimodal.py,sha256=
|
164
|
+
sglang/srt/layers/logits_processor.py,sha256=uduvaN1pFHX3imlx8FbmIG6aJMniOoPmoP0AGimQn5s,26630
|
165
|
+
sglang/srt/layers/multimodal.py,sha256=wVo5-OLz-J4k8fM0ld1ZKMtzOsSue4aa-uILak-I8Bk,2398
|
161
166
|
sglang/srt/layers/parameter.py,sha256=jCg0G-12GZqTa9gGOqEtVCsnnuBGnYg0VmF6pc6oCFg,17455
|
162
167
|
sglang/srt/layers/pooler.py,sha256=uZ6WX1FLMEafZwusyZdm6KuVlIwSjbKrdwk2qzgqNGk,3812
|
163
168
|
sglang/srt/layers/radix_attention.py,sha256=vtzCPfEiWYunglcLfqO9dcATb-LXzU4sFBwIWsWHbiY,3513
|
@@ -175,13 +180,14 @@ sglang/srt/layers/attention/flashattention_backend.py,sha256=rYsBAsFS1qNNe-3CVIa
|
|
175
180
|
sglang/srt/layers/attention/flashinfer_backend.py,sha256=HQp14R29hQVXYjATKPoGaHY9LJ9e1cw2k-RhBEmFxt8,51505
|
176
181
|
sglang/srt/layers/attention/flashinfer_mla_backend.py,sha256=NirRlNFxD0S2EncvvmrerIxNfuTnYpvemZV3MR5_FnU,34040
|
177
182
|
sglang/srt/layers/attention/flashmla_backend.py,sha256=5iSic5ho-lkXNas9mR3uLbXbEl-do31gc8gjR7-a79k,20711
|
183
|
+
sglang/srt/layers/attention/hybrid_attn_backend.py,sha256=ZIzBeR5r0SMkNLekEWPUCas_TCNOmzZ725YaueYVUlM,3196
|
178
184
|
sglang/srt/layers/attention/intel_amx_backend.py,sha256=hH9lxXs_o66_IVs58oPhRCOXGUrHbXSz82FKtHr9GO8,4092
|
179
185
|
sglang/srt/layers/attention/merge_state.py,sha256=OnWmNV6LYTUj0y8SCy_IK7A9q6v9bZwR4_XiLOIvbrk,1423
|
180
186
|
sglang/srt/layers/attention/tbo_backend.py,sha256=CBnSDUmDvT-Tbhi9NHNo1LLkgUAcIsWMJgKa-sAIKW8,9318
|
181
187
|
sglang/srt/layers/attention/torch_native_backend.py,sha256=RsI4lYiQJAj-Xy0vKkzbyKOWHAVPOD9b0n0dEehwlos,9496
|
182
188
|
sglang/srt/layers/attention/triton_backend.py,sha256=AryrPc4kzioxkiucPkApMA__0HoQTKOHNpwN0GEgm98,36923
|
183
189
|
sglang/srt/layers/attention/utils.py,sha256=KiRSIHaGBhErUhPWCSD3Uiaqb0pSEF0NJFwTTQXoMks,2887
|
184
|
-
sglang/srt/layers/attention/vision.py,sha256
|
190
|
+
sglang/srt/layers/attention/vision.py,sha256=zI1Pq2Tx4SwOzYhIarxNtlTiZeVMWxVf8j-Pf7xhzCA,18111
|
185
191
|
sglang/srt/layers/attention/triton_ops/decode_attention.py,sha256=eN0Lqpu5h27lMPuRnisX4jrU0WR061pCFSrh3vLlSD0,18936
|
186
192
|
sglang/srt/layers/attention/triton_ops/double_sparsity_attention.py,sha256=FbgrPO5GvEGEvojV7Ss0jX7384e9vuzX2goDCQuP1Zk,31124
|
187
193
|
sglang/srt/layers/attention/triton_ops/extend_attention.py,sha256=ImpCuxszoeVdr6wzaLR3OGW3-50P9EAEvsRCdUVIrxE,14261
|
@@ -193,15 +199,15 @@ sglang/srt/layers/moe/cutlass_moe_params.py,sha256=9NRCmgP_Ug3gGqCcpi-x-QRbLjCNp
|
|
193
199
|
sglang/srt/layers/moe/cutlass_w4a8_moe.py,sha256=bc8s5Oc36pCF0VozxZuqFOOU2fov-4wGPlyduye3Imk,7296
|
194
200
|
sglang/srt/layers/moe/fused_moe_native.py,sha256=SkOFgU23aZfHmLhnkVhdUWGx0-2wbo6uu9ohKBp2yOI,3486
|
195
201
|
sglang/srt/layers/moe/router.py,sha256=UrPieRvemN7Ew48gtG7DA2xhNDBRSnZxzugTEBI-0_E,12006
|
196
|
-
sglang/srt/layers/moe/topk.py,sha256=
|
202
|
+
sglang/srt/layers/moe/topk.py,sha256=ttPFvXzTY0eyh6UH8fJZCjyDp-R9EssSP8m229MxQ_4,26925
|
197
203
|
sglang/srt/layers/moe/ep_moe/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
198
204
|
sglang/srt/layers/moe/ep_moe/kernels.py,sha256=UFH-XacP3aBjAcn6_SeC641jGzjSrY0fz50IB9URix4,42706
|
199
|
-
sglang/srt/layers/moe/ep_moe/layer.py,sha256=
|
200
|
-
sglang/srt/layers/moe/ep_moe/token_dispatcher.py,sha256=
|
205
|
+
sglang/srt/layers/moe/ep_moe/layer.py,sha256=7VyB__QLkEJg-WLwxZ8yv9Qc3CFDw5HvmNlkKmQ1JFY,49501
|
206
|
+
sglang/srt/layers/moe/ep_moe/token_dispatcher.py,sha256=CJhFTtOydsLSXMYNfzAASPI5p_oP7c4krGwX4bBcT1s,22909
|
201
207
|
sglang/srt/layers/moe/fused_moe_triton/__init__.py,sha256=4NXZHbCw-G-uSnNUj4up0yh3xBDPnT-x0pdoIr0lku8,831
|
202
208
|
sglang/srt/layers/moe/fused_moe_triton/fused_moe.py,sha256=6NcY-6UHQUSVP4qzc8jis4h-W9BEV4lnGmrJMx2S6eA,57164
|
203
|
-
sglang/srt/layers/moe/fused_moe_triton/layer.py,sha256=
|
204
|
-
sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py,sha256=
|
209
|
+
sglang/srt/layers/moe/fused_moe_triton/layer.py,sha256=U3Qaffyj-aGFc6TvlB057eCbiv9oHbPDdwPVHsM7FJs,23278
|
210
|
+
sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py,sha256=HOEJJyehKTUIjkcsTc29aEXHqVLgDlv_jQL5fjvuUxU,5856
|
205
211
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_1_0/E=1,N=14336,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json",sha256=iNGsE2ZeVnQEnN4A8UJ9Jv0d3hbRF2MJ9oBgjup5Szk,2737
|
206
212
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_1_0/E=1,N=14336,device_name=NVIDIA_A100-SXM4-80GB.json",sha256=JJN0hryyLr5Zv3dSS7C8cPFhAwTT6XxUVnBGMZvV6JA,2752
|
207
213
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_1_0/E=1,N=1792,device_name=NVIDIA_A100-SXM4-80GB,dtype=int8_w8a16.json",sha256=ouRyZ5PEMPP2njPftCNhs-1g1y6wueWLmhI7G1SjV1k,4131
|
@@ -346,6 +352,7 @@ sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py,sha256=hoWVdf8ry8ID
|
|
346
352
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=128,N=768,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=pdQ1RvXvdWDn8Y8-8MAX3vn-T-wbtkZvHV9GZZvNjnc,3266
|
347
353
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=128,N=768,device_name=NVIDIA_H200.json",sha256=JraM-Nvbg5V_TJkSl6UPFYZN1zHHoIbr2pAcksenoTY,3248
|
348
354
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=128,N=96,device_name=NVIDIA_H20.json",sha256=JtcHRlPz8xQEAqJ9EWI63oYvdmjQFG6VTHqtt85VOSA,3221
|
355
|
+
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=160,N=320,device_name=NVIDIA_A800-SXM4-80GB,dtype=int8_w8a8.json",sha256=ZRON2XIRRZrvJ3tbPuQ7-0HqGZ_FkPpRN3yj1isGmpg,3255
|
349
356
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=257,N=128,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=ttFuL1rKnEpShItmlGyClOfpUwLi5cZYFaVgo7CAxnA,3268
|
350
357
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=257,N=128,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=Pc2ri-qDByslfD8LgJtRB9QDWjI-KqMeA5ftltiwy8k,3262
|
351
358
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=257,N=256,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=uuua9V6-1H38ASMPQ60t7UOtV0IQLZWBCl13MwQ25Ss,3264
|
@@ -362,7 +369,9 @@ sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py,sha256=hoWVdf8ry8ID
|
|
362
369
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=288,N=64,device_name=NVIDIA_A800-SXM4-80GB.json",sha256=3T8_rF2PEojhgTMyQ8DscXgJCWWdWfDPj4M434zWcA4,3243
|
363
370
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_2_0/E=8,N=7168,device_name=NVIDIA_H100_80GB_HBM3.json",sha256=tme0ydWzIxdABZLk4tU8G_X2dJUYGGZNkQzNGcmcvUc,3261
|
364
371
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=128,N=384,device_name=NVIDIA_H100_80GB_HBM3,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=LNx1W_BsWZcpHomiScCRap46dV0-F7S_w3Htskoqlm8,3263
|
372
|
+
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=160,N=192,device_name=NVIDIA_H200,dtype=fp8_w8a8.json",sha256=rqPWpRfa9s7USa9cUyHAVP50WYkCqXDJzfKZhqk6GHw,3267
|
365
373
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=160,N=320,device_name=NVIDIA_H20-3e.json",sha256=cWkEpNqnyn0QS8HcgWiwWI1xqh_U93_S1kyNeb69aOc,3238
|
374
|
+
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=160,N=384,device_name=NVIDIA_H200,dtype=fp8_w8a8.json",sha256=jGWfZNP5T1en8KX0_94lqqTKRQXC_Kkq_lJrtY4gE8Y,3263
|
366
375
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=257,N=256,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=fo1akCuGoqcGwQgfh56hApgg-wLXfo9kHHksE_6m1F4,3262
|
367
376
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=384,N=128,device_name=NVIDIA_H20,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=ENDYnC1ljPNpDVHdmqR7UbgDWT1IP_55ZIQ_Oae34mI,3264
|
368
377
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=384,N=128,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=5UAs8Z1myw8F-5FY2HHWd5WMT2ii9X-aYIYkf3FsA8s,3269
|
@@ -370,28 +379,31 @@ sglang/srt/layers/moe/fused_moe_triton/triton_kernels_moe.py,sha256=hoWVdf8ry8ID
|
|
370
379
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=385,N=128,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=rgF-KSAUXmI7QGrhMUywPVfeTxz_j8_IP4N-6bkYdak,3260
|
371
380
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=385,N=128,device_name=NVIDIA_H200,dtype=fp8_w8a8,block_shape=[128, 128].json",sha256=7aaoi0MI2xC5ZJkflkTZAH4tawTPdmUpmcBOgOInKxs,3268
|
372
381
|
"sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_3_1/E=8,N=7168,device_name=NVIDIA_H100_80GB_HBM3.json",sha256=a9XJZ0XudUTebAdN5guVq-3XSvZetHIRjatO1-_0xCk,3257
|
382
|
+
sglang/srt/layers/moe/token_dispatcher/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
383
|
+
sglang/srt/layers/moe/token_dispatcher/base_dispatcher.py,sha256=Ut-B7CmWcVPZCPodqBobIdfbclmgs03J3A-zAU3nyNQ,1095
|
384
|
+
sglang/srt/layers/moe/token_dispatcher/standard.py,sha256=jseBfZpst5iCDxXxQV_-53qb0VMqy8RHqXJMf86Bdgg,430
|
373
385
|
sglang/srt/layers/quantization/__init__.py,sha256=lMjOmhMntbY55RR5eKKLB338tdzlxGuBx-pT4f9jmIM,8130
|
374
386
|
sglang/srt/layers/quantization/awq.py,sha256=gLEOx6QoGQ0McBEURXPVidKGsdNUjqc9sJwGLQx86rE,26690
|
375
387
|
sglang/srt/layers/quantization/awq_triton.py,sha256=4XwJbRNlFmlvjbMGtWF1Wvk26Kf96162vNEmuI5Kars,11737
|
376
388
|
sglang/srt/layers/quantization/base_config.py,sha256=CiSxG5iuOa-12V9kIOh0Fu8Y7knZxfFbbNGTo6nEaRE,7173
|
377
389
|
sglang/srt/layers/quantization/blockwise_int8.py,sha256=3avetVUBjCiuYWBJQbElbVXWpx0SgsFsDc5F9F_zHgA,13879
|
378
|
-
sglang/srt/layers/quantization/fp8.py,sha256=
|
379
|
-
sglang/srt/layers/quantization/fp8_kernel.py,sha256=
|
390
|
+
sglang/srt/layers/quantization/fp8.py,sha256=sIA-PCDDu2-AtcwvPcvzitxjCtrvEx_dig_AoocCfhY,47100
|
391
|
+
sglang/srt/layers/quantization/fp8_kernel.py,sha256=8alryD9sEROh9Dh4cdPrGIpuOPbT6Hx6PNSe8fJc9zM,40856
|
380
392
|
sglang/srt/layers/quantization/fp8_utils.py,sha256=6FVR6_Y8akP4ihCaGG_bLYurcC_PShPtiPgiEvGN5bE,25605
|
381
393
|
sglang/srt/layers/quantization/gptq.py,sha256=3q9xfq5sJJHcrVr9C8-l9JdcLFvwKAAfWJEfZW6CzLs,39145
|
382
394
|
sglang/srt/layers/quantization/int8_kernel.py,sha256=GfRC9FOn9exNvK4QHbUeBj3Hhv32VcyGphapFPt5b84,12625
|
383
395
|
sglang/srt/layers/quantization/int8_utils.py,sha256=YK9CS-lb_n91kNCTKK5o5apYF31V2giDg5G5VKrpcUA,2356
|
384
396
|
sglang/srt/layers/quantization/kv_cache.py,sha256=_9pF5rwvB7ta6Gdc5YKVVGbNzYwqmhIx4TrX1-xnodQ,3261
|
385
397
|
sglang/srt/layers/quantization/marlin_utils.py,sha256=ShsoxcVTYo1Es3XKp9mlQNw-TsRF_o4mHIO5MfAP5HA,25906
|
386
|
-
sglang/srt/layers/quantization/modelopt_quant.py,sha256=
|
398
|
+
sglang/srt/layers/quantization/modelopt_quant.py,sha256=oj6jJ66s66kMCufuJapUxGaAX8PjuJB2pdzZuUDZcrM,38357
|
387
399
|
sglang/srt/layers/quantization/moe_wna16.py,sha256=NGcFjiMXwfOX10dkHAY3EgZHxox2k7I402Bvwj3jW1Q,19105
|
388
400
|
sglang/srt/layers/quantization/petit.py,sha256=74Jn0eICq23v_b4rEctDaYTINtbqkdM6IXycT-_FbbI,8954
|
389
401
|
sglang/srt/layers/quantization/petit_utils.py,sha256=-gy4zMhqNoIA1R0n7-5C0efV54jHonCUgPDUUFjGsyM,3245
|
390
402
|
sglang/srt/layers/quantization/qoq.py,sha256=jM96uJfKz3vo1B6SSIrQXNT3vnD3UJrepK2eY-tSQU4,8139
|
391
403
|
sglang/srt/layers/quantization/scalar_type.py,sha256=nBk4THjeRWMGW3hKWRdrjfQ0gwpuBxmYqapDBaRpnr4,12395
|
392
|
-
sglang/srt/layers/quantization/unquant.py,sha256=
|
404
|
+
sglang/srt/layers/quantization/unquant.py,sha256=zmxzCg46IOjIILRPFaZ4FPxodVT83X5nrFAKlkYUiVI,12330
|
393
405
|
sglang/srt/layers/quantization/utils.py,sha256=9ZmfS0MMqueQ6zIxMIFoQy2OFBQ5vHjOHBgVvUn7ews,16213
|
394
|
-
sglang/srt/layers/quantization/w4afp8.py,sha256=
|
406
|
+
sglang/srt/layers/quantization/w4afp8.py,sha256=cYLUuDARPP6NsiYSihDnJYZwLakDL-RGYokTIJ-v_2I,11074
|
395
407
|
sglang/srt/layers/quantization/w8a8_fp8.py,sha256=I_W9A4Pnluaf-SaT8zq-tAJTt1acpxfE_kG5mqdc5cE,10250
|
396
408
|
sglang/srt/layers/quantization/w8a8_int8.py,sha256=VGrVELArNuyw4FEtUefaeK8ayuX7cpXj1OOqCmOCzm4,34902
|
397
409
|
sglang/srt/layers/quantization/compressed_tensors/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
@@ -562,7 +574,7 @@ sglang/srt/lora/layers.py,sha256=FaTYf0w3zbKKKAT1nJdAzGRtxvySB6FpNKXvEi5RHH0,129
|
|
562
574
|
sglang/srt/lora/lora.py,sha256=sRoU7BdlTV4cnBv1-u3QjisVR9XQrrqAsmhxdNtd8dU,9433
|
563
575
|
sglang/srt/lora/lora_config.py,sha256=qDgMTx_69jyJUl29O5FxLzYa0BMhqYVXWXfyyVOvGm0,1684
|
564
576
|
sglang/srt/lora/lora_manager.py,sha256=ABuqg1IhY9LNXgJd_v3Yx_VHS52t4SapHmhLiH1thdc,20147
|
565
|
-
sglang/srt/lora/lora_registry.py,sha256=
|
577
|
+
sglang/srt/lora/lora_registry.py,sha256=z3Ib7nlr1gajeSwJjC7sL8eMiaLiZo7dOB3WZnlyBVM,7676
|
566
578
|
sglang/srt/lora/mem_pool.py,sha256=UYvTgEPvBoVsaX5rEnipebOAlo75pML6_NO9GfCFzl8,12148
|
567
579
|
sglang/srt/lora/utils.py,sha256=5G0KiDEn-Zg4OgIOjWnU2ZdQGNstZswoAog9xU4GVDA,5660
|
568
580
|
sglang/srt/lora/backend/base_backend.py,sha256=EIz8I-GIrdmK4fISw3ENhbJVVITaxKfyLxHXGPU4fPs,5044
|
@@ -573,21 +585,22 @@ sglang/srt/lora/triton_ops/gate_up_lora_b.py,sha256=QURCYxHNR8Ls4SQtt3dvdgjvdDVh
|
|
573
585
|
sglang/srt/lora/triton_ops/qkv_lora_b.py,sha256=iz5scFNo2YFTeiM8beWg9Z1oZI-6AM_T1wBMCQ6qp2Q,6485
|
574
586
|
sglang/srt/lora/triton_ops/sgemm_lora_a.py,sha256=Ksova04wgeGsFqGOXWqJtMYaHgyUYcx8VU42BZQOkVA,5129
|
575
587
|
sglang/srt/lora/triton_ops/sgemm_lora_b.py,sha256=VqCAFvUtq_l-0RGIkx3W_fzD55QcW2FAcqpvSHOBFp4,5115
|
576
|
-
sglang/srt/managers/cache_controller.py,sha256
|
588
|
+
sglang/srt/managers/cache_controller.py,sha256=HoI6_os2h8rrvyEJ6Kk4ibUF3aKH-9Z8K5pksc6BECQ,24603
|
577
589
|
sglang/srt/managers/configure_logging.py,sha256=8sNXZ2z9pBWOwn-X3wyz013Ob8Nbm1zDxRkxoZjH-l4,1633
|
578
|
-
sglang/srt/managers/data_parallel_controller.py,sha256=
|
590
|
+
sglang/srt/managers/data_parallel_controller.py,sha256=gcg_TGXJCcdh2GFT0Fu-CL_Ihc76jKrMH3LXfZlMjtk,12289
|
579
591
|
sglang/srt/managers/detokenizer_manager.py,sha256=SpLxTsSPKBZfD-ZMhJ5zpPPGuUb8PmcYgFSL9CsurU4,10696
|
580
|
-
sglang/srt/managers/io_struct.py,sha256=
|
581
|
-
sglang/srt/managers/mm_utils.py,sha256=
|
582
|
-
sglang/srt/managers/multimodal_processor.py,sha256=
|
583
|
-
sglang/srt/managers/schedule_batch.py,sha256=
|
592
|
+
sglang/srt/managers/io_struct.py,sha256=_wMPKcCynoq6-rDOkcKXFOUjSXUzIEmTHES8760jbdI,38058
|
593
|
+
sglang/srt/managers/mm_utils.py,sha256=VrJiv0k_zDa83BsfVv8rmckgXKsvF0Wz88bCAb6R1as,29601
|
594
|
+
sglang/srt/managers/multimodal_processor.py,sha256=mLnGcUHIvjruwoVhqiz4j7RMYMsZnvQMH8JnKCITDlY,1805
|
595
|
+
sglang/srt/managers/schedule_batch.py,sha256=xmWSPAsllmPaWtKNrf-P8IX6fuQMTD-vZuk1y01vUjA,76289
|
584
596
|
sglang/srt/managers/schedule_policy.py,sha256=BFYItMXtrsO5157tpFhYwLpZ8NSlzR4_gNZC1iE10D4,22400
|
585
|
-
sglang/srt/managers/scheduler.py,sha256=
|
597
|
+
sglang/srt/managers/scheduler.py,sha256=AJAbKjP6rBePpb4iVoD9X-99z6-cOaxUBS3OUlXA92U,123741
|
598
|
+
sglang/srt/managers/scheduler_input_blocker.py,sha256=zP8xU_UmU2H0AB6sEqvivDcDB1QDgTauNDYDIJ7Nez4,3683
|
586
599
|
sglang/srt/managers/scheduler_output_processor_mixin.py,sha256=Ei-9lyfrSQAlL9_ULuJLAhA7qpvUzjsBDzBBTDdnt4Q,31394
|
587
600
|
sglang/srt/managers/session_controller.py,sha256=dzlMNZlo20FTSl64QqK7y7pElsdCy8ICOWWBPTBVwgs,6040
|
588
601
|
sglang/srt/managers/template_manager.py,sha256=RrwRA2oqId_PMQ98qJQGwIxMroOxiorl2sGC9ARou_0,8543
|
589
|
-
sglang/srt/managers/tokenizer_manager.py,sha256=
|
590
|
-
sglang/srt/managers/tp_worker.py,sha256=
|
602
|
+
sglang/srt/managers/tokenizer_manager.py,sha256=tq0tOSTJLZF8AE1iXzG-NrJyS8PHmmgQFXFlYGvZFUM,78025
|
603
|
+
sglang/srt/managers/tp_worker.py,sha256=PKNYIkInFkCITimwXxvXwDcHEwTgfQUfOZjXFvRKo5I,11617
|
591
604
|
sglang/srt/managers/tp_worker_overlap_thread.py,sha256=kK0pn8mz-XgLGvVlJK6LvDVp4lnFk5vJH-Cs9NQ5Un4,10959
|
592
605
|
sglang/srt/managers/utils.py,sha256=9_VGE48EK0PXVJ26aYvbRJ6n7gIZALvCcf6uZCccCgM,1369
|
593
606
|
sglang/srt/mem_cache/allocator.py,sha256=id4riucsG9d-P1ikj3-DLJbzhM8-MuJfvDmB8H4g9-o,23364
|
@@ -595,7 +608,7 @@ sglang/srt/mem_cache/base_prefix_cache.py,sha256=hLS2ncTMAz7Kpdk5pNwn5c6g8b61_K9
|
|
595
608
|
sglang/srt/mem_cache/chunk_cache.py,sha256=FU_fq8e_dodgU8EkRqE4Jr0sKwCXQnKmapO88k6_JfU,3207
|
596
609
|
sglang/srt/mem_cache/flush_cache.py,sha256=GYcxmNXh4hsMpFfNOuCTpKilW7guZwTtAg_usVeM3J0,979
|
597
610
|
sglang/srt/mem_cache/hicache_storage.py,sha256=pvNeooVmCZBeZFgSHWNGPZzqtHI_qJ5BpTy_A4i6gg8,5539
|
598
|
-
sglang/srt/mem_cache/hiradix_cache.py,sha256=
|
611
|
+
sglang/srt/mem_cache/hiradix_cache.py,sha256=3BXlIZUyQk2dkwuwsMvSUSXd8fCDLPvcTLOp9FnzxBY,24779
|
599
612
|
sglang/srt/mem_cache/memory_pool.py,sha256=jQdWgOr6hpMLbGaEJc4immBgZG_MZbx8LyOv029ASVw,40932
|
600
613
|
sglang/srt/mem_cache/memory_pool_host.py,sha256=xcjYyC94FOH_NguWbLUG4vF_CSBcxA_msuRsBasYCPo,10387
|
601
614
|
sglang/srt/mem_cache/multimodal_cache.py,sha256=wZl2KeEl3xeoEsYdH33UoM-FO8kqfLo_XUgereJVvoM,1348
|
@@ -605,7 +618,7 @@ sglang/srt/metrics/collector.py,sha256=FH6Wil2lMK7d7q74CfxfdULCaHKDXeAoo8UBZqqde
|
|
605
618
|
sglang/srt/metrics/func_timer.py,sha256=VFyNRrbnKVCwnQsrlLin1lITJfjQpf9m8sGPqL5LIsQ,3438
|
606
619
|
sglang/srt/model_executor/cuda_graph_runner.py,sha256=6njRGysrhBDi2bAt5DyjGnyrexYvuEVsAJwQBBW8Ut0,31886
|
607
620
|
sglang/srt/model_executor/forward_batch_info.py,sha256=NvUhReevoy6UfO4991P_5MPQyfefC3BOoTwTqUwMbYk,38430
|
608
|
-
sglang/srt/model_executor/model_runner.py,sha256=
|
621
|
+
sglang/srt/model_executor/model_runner.py,sha256=jC2Int_-_LQMYxQHfnXa76IMC7anMAbWpqgcsJweh28,72587
|
609
622
|
sglang/srt/model_loader/__init__.py,sha256=zGZkOBz1zx-pkaIy47BasL3fjDlAcxAXUTjInOhXHAE,919
|
610
623
|
sglang/srt/model_loader/loader.py,sha256=wxSUYrqv23w9yHYy9Lm0ozWggs9pf6Co4yqe1UO8PZk,59799
|
611
624
|
sglang/srt/model_loader/utils.py,sha256=svSrPQxViBz_-0_sByu4coCYrzJ51-4FCi_bFqa6rDo,4474
|
@@ -619,7 +632,7 @@ sglang/srt/models/dbrx.py,sha256=4pn_fdoATg01VEqNnIAxNEsKV5XU7gwHyd289eydq1s,155
|
|
619
632
|
sglang/srt/models/deepseek.py,sha256=sqobTr9a9c5pNLNJZE-WrJIE_qBq0lC5gCGv9NpGU_Q,17364
|
620
633
|
sglang/srt/models/deepseek_janus_pro.py,sha256=td8xGs6ARfJ8AQCYwUhMOZoWigrAs7m3trF5-kXCqik,70418
|
621
634
|
sglang/srt/models/deepseek_nextn.py,sha256=47fehxRdiOizr0rdLg5f1fzQEx6gGAOcDcWKtblloyk,5928
|
622
|
-
sglang/srt/models/deepseek_v2.py,sha256=
|
635
|
+
sglang/srt/models/deepseek_v2.py,sha256=S2KRJczyfTZ6Yl3PqF_2GAL7yUXHPl2BDhnm07JA_UE,103547
|
623
636
|
sglang/srt/models/deepseek_vl2.py,sha256=7X3gI4DE7guUtNJvMLf06jO8MUHKa2Aan8evZg5hsXc,13061
|
624
637
|
sglang/srt/models/exaone.py,sha256=TpO-rtCpEZ8Ua7hGFnS8l2oAYhY0Pij50grc9WQ2mvc,13576
|
625
638
|
sglang/srt/models/gemma.py,sha256=2atKwFN5wCsJCmBRmURX9vaCcFjxlFirL7xsnG5KPlI,14180
|
@@ -631,6 +644,8 @@ sglang/srt/models/gemma3n_audio.py,sha256=isgKfjA5UieYawxU6medL2ssXlzYPqAbagDBnL
|
|
631
644
|
sglang/srt/models/gemma3n_causal.py,sha256=kBDlpSWAg6ZDmUPe08Q7sj5CbdVD_0h07zwJw8sKu7I,36335
|
632
645
|
sglang/srt/models/gemma3n_mm.py,sha256=pL_fvnkz9fKaC2Afms5KG3Me3u5sp1C2-mAUhqPGrQ4,18701
|
633
646
|
sglang/srt/models/glm4.py,sha256=2VQzUqFkQTy_2nfkxP9SF6_9kKLTZUExGRjge7r99Es,11265
|
647
|
+
sglang/srt/models/glm4_moe.py,sha256=vNJfab2DVhMEU-9RsPd7zyLofkX7SqYHRY97vFcYkDA,39864
|
648
|
+
sglang/srt/models/glm4_moe_nextn.py,sha256=cWB6o3w7xBvgi2ouGKsqZ3iovKH65QiRQul42wcgzMk,5897
|
634
649
|
sglang/srt/models/gpt2.py,sha256=kclhxEs8oJk1KCyhmAqo7rZqecVGGHYkc-a1WZi3aIk,9841
|
635
650
|
sglang/srt/models/gpt_bigcode.py,sha256=1D6bi8Zu760gCRZkvdLHFcg8kCkY35ARwQYaMDtYhl4,10307
|
636
651
|
sglang/srt/models/granite.py,sha256=5WOJyNYAlt5RNHSexNfPNihhSxIMd7wPzju1cTixKig,20852
|
@@ -640,7 +655,8 @@ sglang/srt/models/hunyuan.py,sha256=GVDeMKgcqW6HMFWPUukznL5JoDW4CpOpS5Q_PXlbHw8,
|
|
640
655
|
sglang/srt/models/idefics2.py,sha256=U3khd3hbdawJeRNXsxmaKHdssOCT5TPOZ1D-2_zHoQo,12079
|
641
656
|
sglang/srt/models/internlm2.py,sha256=F_iNY1gUqzAjAuUatcE47gnrcoTh5_08PY2Rw9tKr9M,13150
|
642
657
|
sglang/srt/models/internlm2_reward.py,sha256=ndfGmyqYZbVZ7C7rJ-v9oK3wa-EpoBGybS8MlyKZi2E,2522
|
643
|
-
sglang/srt/models/
|
658
|
+
sglang/srt/models/interns1.py,sha256=nResXUzfcHZoJie1gHuYwyVir7Ys2fWp3hjzTs6Utvk,14182
|
659
|
+
sglang/srt/models/internvl.py,sha256=o6XSAvnMYK5tTPbln9RI9BfRfAQZBE8_xmGPffrznnQ,27838
|
644
660
|
sglang/srt/models/kimi_vl.py,sha256=XulRiufQzOSWZe8-bUl2OEdznvsVU9OBsEYipaX3YkM,12877
|
645
661
|
sglang/srt/models/kimi_vl_moonvit.py,sha256=M5L7AJOJ2Zh2pqLZAq8aRqhbTSlipr8XOoD3ix6o2sU,23908
|
646
662
|
sglang/srt/models/llama.py,sha256=Nsxv2s1sOmLnZsTHkwbTh53cUFEQ2fnKDQi5BFbFcsk,28994
|
@@ -650,13 +666,13 @@ sglang/srt/models/llama_eagle.py,sha256=Ubh_jmtfiOnriwvHgQT0ZGID7JvYvdSi_QGHOIkE
|
|
650
666
|
sglang/srt/models/llama_eagle3.py,sha256=715soylDE3kOMWskbPy_70BFxaSc5ptaHCS1v3-ViA0,8579
|
651
667
|
sglang/srt/models/llama_embedding.py,sha256=zq-_lNu35VBFc7eemiam0zdkGIE8fzrgk5OWYfirZnA,3254
|
652
668
|
sglang/srt/models/llama_reward.py,sha256=LF2nqMV5XOrljGjAwJg43mBv3z6Q040I2EYlgZeCp8k,4681
|
653
|
-
sglang/srt/models/llava.py,sha256=
|
669
|
+
sglang/srt/models/llava.py,sha256=xzYip_BAwpzSIdZre43LZiyTpFISa0ZCLdO6LUSbaCg,37702
|
654
670
|
sglang/srt/models/llavavid.py,sha256=-CSk0RJ2MQeb81sh-RISeVJFaI-XWY6nR6_I594MkME,12818
|
655
671
|
sglang/srt/models/mimo.py,sha256=Mp-iFp4YHuiuq-H8enUF5K5QbMnVcvEa6mURH6vM3yM,6140
|
656
672
|
sglang/srt/models/mimo_mtp.py,sha256=jSmqJAXu7G3OO7jW1oa2suI4H_Yl7u5ZT7w4lHFbHhg,7292
|
657
673
|
sglang/srt/models/minicpm.py,sha256=CzBJyZtfMpp8jvlEl29DHI7HLVq-CxuqP1UHwxIbaUI,14567
|
658
674
|
sglang/srt/models/minicpm3.py,sha256=6-gfHSfXwyB5zw7AIj-c5TzjpEKR6CM0H67MrD8LVUE,19347
|
659
|
-
sglang/srt/models/minicpmo.py,sha256=
|
675
|
+
sglang/srt/models/minicpmo.py,sha256=st8I13Ere5RkRHHQQN1LgZh1ybUA3_buxaaE4XlV8ag,76123
|
660
676
|
sglang/srt/models/minicpmv.py,sha256=g4aAOG8PVGkZqGmlneVaScAtNpSLqSVbZGRAikyrBy8,30394
|
661
677
|
sglang/srt/models/mistral.py,sha256=ni7PppA_5X4ivTnIps3yoEaMEL-G6Gaf9b5hGC1vU5Q,3475
|
662
678
|
sglang/srt/models/mixtral.py,sha256=j2whKt4GdQqtpKL5iN0Vb4TO4el1E2qEpUCI57E_47Q,17222
|
@@ -680,11 +696,11 @@ sglang/srt/models/qwen2_5_vl.py,sha256=vOuIyK_NLuGgz9LlAdyM7cSQUU93VnBCYXUr_n4iO
|
|
680
696
|
sglang/srt/models/qwen2_audio.py,sha256=lVeYmC29eVZIR9I6ZjFe4xvC0wSXR0NtGCli6fSXtms,7870
|
681
697
|
sglang/srt/models/qwen2_classification.py,sha256=dGrMm4ebd30_lBhHOhaV57ig2iOTx3nqB4GEzsrRIM8,2747
|
682
698
|
sglang/srt/models/qwen2_eagle.py,sha256=6spFg6JYQGeUXLv2Q5eMkVzZvnqGyvW6SdVBtFMxlTM,4806
|
683
|
-
sglang/srt/models/qwen2_moe.py,sha256=
|
699
|
+
sglang/srt/models/qwen2_moe.py,sha256=drTwlJlP0PLDl-Rl2h4GaOmOxexV1SMjMP97izhVpLk,25822
|
684
700
|
sglang/srt/models/qwen2_rm.py,sha256=-mQXDEv11p-I1HXgYLTtY6ROem6UYorO958WsDrzsgs,2837
|
685
701
|
sglang/srt/models/qwen2_vl.py,sha256=ZRAL32Ymt4ZCBqB0JorXAm4ZBrECZ9EbaXKaTSikkQM,21366
|
686
702
|
sglang/srt/models/qwen3.py,sha256=COSC1TsCQNTq3E1QoVJYj1yTn-CFEVEX7keRwrrBXLs,20121
|
687
|
-
sglang/srt/models/qwen3_moe.py,sha256=
|
703
|
+
sglang/srt/models/qwen3_moe.py,sha256=XJ-93F8gVWV5u8MsvC0DKwIDqSBgORO60l8Vd6T0dCE,31541
|
688
704
|
sglang/srt/models/registry.py,sha256=bH9H0OiNV0Cn7FRBM30Oh6dxICTs9mJscdLtHYPJvxc,3783
|
689
705
|
sglang/srt/models/roberta.py,sha256=3k53V2Gbezk3jU-D03thXx8csGn5DxFK01ZQ9WFiWPI,9828
|
690
706
|
sglang/srt/models/siglip.py,sha256=DIVJRwdtpLD2QT1kVPIHw7Bn9BE40xJbkHujDNzkjgA,9892
|
@@ -696,23 +712,23 @@ sglang/srt/models/xverse.py,sha256=DsNVI9JpzN4jj0Ry6aTrj7r-xq5YLOoDX2kH4YLJA-I,1
|
|
696
712
|
sglang/srt/models/xverse_moe.py,sha256=7KCM2-j12towDMNvXkuuYiBOmNauH6NG4Ip40x0khqA,16782
|
697
713
|
sglang/srt/models/yivl.py,sha256=oToK7-u5IGO7xwpJIQ7VtudlK6-zPqJX4bt6_wv0SH8,4850
|
698
714
|
sglang/srt/multimodal/mm_utils.py,sha256=6bV1_a7Iet9W0y3qdd0E-hFM78zalIzNZCYbLzJX2y8,12640
|
699
|
-
sglang/srt/multimodal/processors/base_processor.py,sha256=
|
700
|
-
sglang/srt/multimodal/processors/clip.py,sha256=
|
701
|
-
sglang/srt/multimodal/processors/deepseek_vl_v2.py,sha256=
|
702
|
-
sglang/srt/multimodal/processors/gemma3.py,sha256=
|
703
|
-
sglang/srt/multimodal/processors/gemma3n.py,sha256=
|
704
|
-
sglang/srt/multimodal/processors/internvl.py,sha256=
|
705
|
-
sglang/srt/multimodal/processors/janus_pro.py,sha256=
|
706
|
-
sglang/srt/multimodal/processors/kimi_vl.py,sha256=
|
707
|
-
sglang/srt/multimodal/processors/llava.py,sha256=
|
708
|
-
sglang/srt/multimodal/processors/minicpm.py,sha256=
|
709
|
-
sglang/srt/multimodal/processors/mlama.py,sha256=
|
710
|
-
sglang/srt/multimodal/processors/mllama4.py,sha256=
|
711
|
-
sglang/srt/multimodal/processors/phi4mm.py,sha256=
|
712
|
-
sglang/srt/multimodal/processors/pixtral.py,sha256=
|
713
|
-
sglang/srt/multimodal/processors/qwen_audio.py,sha256=
|
714
|
-
sglang/srt/multimodal/processors/qwen_vl.py,sha256=
|
715
|
-
sglang/srt/multimodal/processors/vila.py,sha256=
|
715
|
+
sglang/srt/multimodal/processors/base_processor.py,sha256=duDeCOOeXRt8lYi34fCe40Gezz0-Ao-rN9Zenx8p4c8,24430
|
716
|
+
sglang/srt/multimodal/processors/clip.py,sha256=5nXZuQYbyf-5mlqPwjhQrlWqxvswImWZQXivChsiHfQ,1057
|
717
|
+
sglang/srt/multimodal/processors/deepseek_vl_v2.py,sha256=Pymag9f0U5L_bBOTFs2q94BbuPJ-VwqSixs3meO0nsM,2569
|
718
|
+
sglang/srt/multimodal/processors/gemma3.py,sha256=7aD6cR60iSC_MqjzdCDMAgtrb_cmy9YJyqLgVRClBT8,1986
|
719
|
+
sglang/srt/multimodal/processors/gemma3n.py,sha256=zml4ovgDzgFZxupnsTfmGtT69ymY_eY0Zp3hR6TkBY0,3056
|
720
|
+
sglang/srt/multimodal/processors/internvl.py,sha256=p8dba1LBEceGObJvPdq3auDhlMKQIXyp_fXM5VMd4BI,9771
|
721
|
+
sglang/srt/multimodal/processors/janus_pro.py,sha256=_4ltWmdtFjbh1VCMRbcgxW61eWIABIzHG-YgwHWaYq8,1471
|
722
|
+
sglang/srt/multimodal/processors/kimi_vl.py,sha256=5i8KmGxnnDtQc7XqUBNyanAfGxhxj8iRFBFONHmpiFg,1565
|
723
|
+
sglang/srt/multimodal/processors/llava.py,sha256=1pb47tlbgH8IUgxTb2YFsK87-x16xmC6BZFHc5r2x9g,8097
|
724
|
+
sglang/srt/multimodal/processors/minicpm.py,sha256=ysJlUaxcBcWjKliz5VmXKVv0ec4QWRHXZk4ga-hXVf0,5354
|
725
|
+
sglang/srt/multimodal/processors/mlama.py,sha256=FgPm2JQzfPFDcGYDm0mqfDfO_eJKy-NiVgF-vId6w8Y,1233
|
726
|
+
sglang/srt/multimodal/processors/mllama4.py,sha256=1YGETKSzMZIa7BOwq3T9cbsYjUcW5sopcpc9JNJ000g,1911
|
727
|
+
sglang/srt/multimodal/processors/phi4mm.py,sha256=8GaUVpGucV4UKoa0XkxN0gDpMZax6UyCiRBX_QaJK78,3857
|
728
|
+
sglang/srt/multimodal/processors/pixtral.py,sha256=Lx5Z-9NkovLDTHlRBznHxGn4r2XhDz8am2XpOskOVQY,3203
|
729
|
+
sglang/srt/multimodal/processors/qwen_audio.py,sha256=jxtArfNuBH4q9LQVEydMDsj85avqBGNl-v1ZZTjC9E0,2367
|
730
|
+
sglang/srt/multimodal/processors/qwen_vl.py,sha256=tZOpz8dabyNvjS4hn0UxLf-EaEZH8kmHjyZaYRm9EZ8,10498
|
731
|
+
sglang/srt/multimodal/processors/vila.py,sha256=vWgV1lTdjiFiTR8w-7fC66-L_tpOOD59c-c9NljGelE,2192
|
716
732
|
sglang/srt/sampling/custom_logit_processor.py,sha256=AwODYVJdRkcQ8PGtJrhzKsqAgn8XZLQbAmR9fGiQzmc,1608
|
717
733
|
sglang/srt/sampling/sampling_batch_info.py,sha256=Duk4Fp99QZyOFld3i4AGwkW9FlULY86CNKv6v0_kA4w,14683
|
718
734
|
sglang/srt/sampling/sampling_params.py,sha256=HbVcCeSXgPW19MRB-v0FTG1D-zHDAY6-toxyndSl-zI,6462
|
@@ -727,11 +743,12 @@ sglang/srt/speculative/eagle_draft_extend_cuda_graph_runner.py,sha256=hLjaD66K3Z
|
|
727
743
|
sglang/srt/speculative/eagle_utils.py,sha256=x5MKA2GkecNAXE58KKhPJRy5KxCLhRkD-C_kVSZIyzc,46433
|
728
744
|
sglang/srt/speculative/eagle_worker.py,sha256=8yPQZxpBvcr--Ooo4nov3LqzkLRL-GOLJlwu91wMAmQ,38650
|
729
745
|
sglang/srt/speculative/spec_info.py,sha256=rhaKG0TzyF9XZYHEWp1jccwTBohSNsUDvxHFtAoOl18,709
|
746
|
+
sglang/srt/weight_sync/utils.py,sha256=jpmfXtiCvUYpwWh4bMgQgv0JSulPi8TwIz4aTsxi0XU,4148
|
730
747
|
sglang/test/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
731
748
|
sglang/test/few_shot_gsm8k.py,sha256=qnEejCyPHGEMBEiNc7T4-t9lvoo2DNN72RXw-ei_TnY,4300
|
732
749
|
sglang/test/few_shot_gsm8k_engine.py,sha256=Yf1Dg0Tk-tY4v1oj_UTxJ7950CWn6kRMbQKDjT2Orgo,3902
|
733
750
|
sglang/test/run_eval.py,sha256=K4Z3OdMPvLeuGJhWWTBD0F-MOUFeBYLvsPnDQ44bd18,4036
|
734
|
-
sglang/test/runners.py,sha256=
|
751
|
+
sglang/test/runners.py,sha256=q01InkXESqxZe19NeH5SzK11HuQG13gFEfm3N7XAIpU,33714
|
735
752
|
sglang/test/send_one.py,sha256=XNZdfZs21tXDBwrDj2pq46WQd5smiKoYETY8U09c1Z8,5079
|
736
753
|
sglang/test/simple_eval_common.py,sha256=OW4Z7O33QqDySRFuHIC4Ouidfup4eaOEuR9i7xJzkDY,12379
|
737
754
|
sglang/test/simple_eval_gpqa.py,sha256=8Xt9Bw05c7SZTYrCZgB68OZUqUbLo69ywiyx0bTvSUk,3220
|
@@ -753,13 +770,13 @@ sglang/test/test_layernorm.py,sha256=2GMWqqNDuGvSMSsEBF5eDCzwVSYA9E6hGhRo6s4ecKg
|
|
753
770
|
sglang/test/test_marlin_moe.py,sha256=PITMGcGF4ci2csXJejdLSNuxpURnFmcflsipAZsGakU,7991
|
754
771
|
sglang/test/test_marlin_utils.py,sha256=gFMJwxbfEEJDeHI-HU5FDF9FFqqcM-yahK0TWyVIVgQ,5171
|
755
772
|
sglang/test/test_programs.py,sha256=DJIbTChloPwaDiEWHQoSTuKv3BGx0-K8Y5e1mDDTACM,18910
|
756
|
-
sglang/test/test_utils.py,sha256=
|
773
|
+
sglang/test/test_utils.py,sha256=672Bp2kighQGrTC4SnIyy8rsoLOxGDjnCtsIhhOBLH4,41171
|
757
774
|
sglang/test/attention/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
|
758
775
|
sglang/test/attention/test_flashattn_backend.py,sha256=_rTG849FwQdVTyGKkqhczaOqngBmRWXFmkl5NnuK1GM,13914
|
759
776
|
sglang/test/attention/test_flashattn_mla_backend.py,sha256=g4O50WblTpM7_Gq2b76k0i25_z01BOUBQ4i6PmyxpO4,10774
|
760
777
|
sglang/test/attention/test_prefix_chunk_info.py,sha256=hpoDe2wfSa6RlUbfyri_c0iyBTb35UXGL9I2Xh6jamM,7772
|
761
|
-
sglang-0.4.9.
|
762
|
-
sglang-0.4.9.
|
763
|
-
sglang-0.4.9.
|
764
|
-
sglang-0.4.9.
|
765
|
-
sglang-0.4.9.
|
778
|
+
sglang-0.4.9.post6.dist-info/licenses/LICENSE,sha256=FJXh51fvTQklojUFY89XVLsjxRcBqOxPs8XNy-2uZ0c,11346
|
779
|
+
sglang-0.4.9.post6.dist-info/METADATA,sha256=CYETYraU2OqIt-j8SUxABCh-qIJAer5LaHRf0NZapvs,27394
|
780
|
+
sglang-0.4.9.post6.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
|
781
|
+
sglang-0.4.9.post6.dist-info/top_level.txt,sha256=yxhh3pYQkcnA7v3Bg889C2jZhvtJdEincysO7PEB09M,7
|
782
|
+
sglang-0.4.9.post6.dist-info/RECORD,,
|
sglang/srt/debug_utils.py
DELETED
@@ -1,74 +0,0 @@
|
|
1
|
-
import os
|
2
|
-
import time
|
3
|
-
from pathlib import Path
|
4
|
-
|
5
|
-
import torch
|
6
|
-
|
7
|
-
from sglang.srt.utils import get_bool_env_var
|
8
|
-
|
9
|
-
|
10
|
-
class _Dumper:
|
11
|
-
"""Utility to dump tensors, which can be useful when comparison checking models.
|
12
|
-
|
13
|
-
Example usage:
|
14
|
-
debug_utils.dumper.dump("layer_start_hidden_states", hidden_states, layer_id=self.layer_id)
|
15
|
-
"""
|
16
|
-
|
17
|
-
def __init__(self):
|
18
|
-
self._enable = get_bool_env_var("SGLANG_DUMPER_ENABLE", "true")
|
19
|
-
self._base_dir = Path(os.environ.get("SGLANG_DUMPER_DIR", "/tmp"))
|
20
|
-
self._enable_write_file = get_bool_env_var("SGLANG_DUMPER_WRITE_FILE", "1")
|
21
|
-
self._partial_name = str(time.time())
|
22
|
-
self.forward_pass_id = None
|
23
|
-
|
24
|
-
def dump(self, name, value, **kwargs):
|
25
|
-
if not self._enable:
|
26
|
-
return
|
27
|
-
|
28
|
-
from sglang.srt.distributed import get_tensor_model_parallel_rank
|
29
|
-
|
30
|
-
rank = get_tensor_model_parallel_rank()
|
31
|
-
full_kwargs = dict(
|
32
|
-
forward_pass_id=self.forward_pass_id,
|
33
|
-
name=name,
|
34
|
-
**kwargs,
|
35
|
-
)
|
36
|
-
full_filename = "___".join(f"{k}={v}" for k, v in full_kwargs.items()) + ".pt"
|
37
|
-
path = (
|
38
|
-
self._base_dir / f"sglang_dump_{self._partial_name}_{rank}" / full_filename
|
39
|
-
)
|
40
|
-
|
41
|
-
sample_value = self._get_sample_value(name, value)
|
42
|
-
|
43
|
-
print(
|
44
|
-
f"[{rank}, {time.time()}] {path} "
|
45
|
-
f"type={type(value)} "
|
46
|
-
f"shape={value.shape if isinstance(value, torch.Tensor) else None} "
|
47
|
-
f"dtype={value.dtype if isinstance(value, torch.Tensor) else None} "
|
48
|
-
f"sample_value={sample_value}"
|
49
|
-
)
|
50
|
-
|
51
|
-
if self._enable_write_file:
|
52
|
-
path.parent.mkdir(parents=True, exist_ok=True)
|
53
|
-
torch.save(value, str(path))
|
54
|
-
|
55
|
-
def _get_sample_value(self, name, value):
|
56
|
-
if value is None:
|
57
|
-
return None
|
58
|
-
|
59
|
-
if isinstance(value, tuple):
|
60
|
-
return [self._get_sample_value(name, x) for x in value]
|
61
|
-
|
62
|
-
if not isinstance(value, torch.Tensor):
|
63
|
-
return None
|
64
|
-
|
65
|
-
if value.numel() < 200:
|
66
|
-
return value
|
67
|
-
|
68
|
-
slices = [
|
69
|
-
slice(0, 5) if dim_size > 200 else slice(None) for dim_size in value.shape
|
70
|
-
]
|
71
|
-
return value[tuple(slices)]
|
72
|
-
|
73
|
-
|
74
|
-
dumper = _Dumper()
|
File without changes
|
File without changes
|
File without changes
|