A high-throughput and memory-efficient inference and serving engine for LLMs
[CI] Reorganize entrypoints CI (#44947)
Signed-off-by: wang.yuqi <yuqi.wang@daocloud.io>
wang.yuqi committed
Jun 9, 2026 at 15:46 UTC
996222f4bffc7b0c1e50bb18b42e9f0f09867d7b
20 files changed
+125
-124
.buildkite/test-amd.yaml
+73
-70
@@ -88,16 +88,16 @@
88
# - Do NOT remove `VLLM_WORKER_MULTIPROC_METHOD=spawn` setting as ROCm requires this for certain models to function. #
89
# * [Transformers Nightly Models]: Whisper needs `VLLM_WORKER_MULTIPROC_METHOD=spawn` to avoid deadlock. #
90
# * [Plugin Tests (2 GPUs)]: #
91
-# - {`pytest -v -s entrypoints/openai/test_oot_registration.py`}: It needs a clean process #
92
-# - {`pytest -v -s models/test_oot_registration.py`}: It needs a clean process #
93
-# - {`pytest -v -s plugins/lora_resolvers`}: Unit tests for in-tree lora resolver plugins #
91
+# - {`pytest -v -s plugins_tests/test_oot_registration_online.py`}: It needs a clean process #
92
+# - {`pytest -v -s plugins_tests/test_oot_registration_offline.py`}: It needs a clean process #
93
+# - {`pytest -v -s plugins_tests/lora_resolvers`}: Unit tests for in-tree lora resolver plugins #
94
# * [LoRA TP (Distributed)]: #
95
# - There is some Tensor Parallelism related processing logic in LoRA that requires multi-GPU testing for validation. #
96
# - {`pytest -v -s -x lora/test_gptoss_tp.py`}: Disabled for now because MXFP4 backend on non-cuda platform doesn't support #
97
# LoRA yet. #
98
# * [Distributed Tests (NxGPUs)(HW-TAG)]: Don't test llama model here, it seems hf implementation is buggy. See: #
99
# https://github.com/vllm-project/vllm/pull/5689 #
100
-# * [Distributed Tests (NxGPUs)(HW-TAG)]: Some old E2E tests were removed in https://github.com/vllm-project/vllm/pull/33293 #
100
+# * [Distributed Tests (NxGPUs)(HW-TAG)]: Some old E2E tests were removed in https://github.com/vllm-project/vllm/pull/33293 #
101
# in favor of new tests in fusions_e2e. We avoid replicating the new jobs in #
102
# this file as it's deprecated. #
103
# #
@@ -617,9 +617,9 @@ steps:
617
- pytest -v -s plugins_tests/test_scheduler_plugins.py
618
- pip install -e ./plugins/vllm_add_dummy_model
619
- pytest -v -s distributed/test_distributed_oot.py
620
- - pytest -v -s entrypoints/openai/chat_completion/test_oot_registration.py
621
- - pytest -v -s models/test_oot_registration.py
622
- - pytest -v -s plugins/lora_resolvers
620
+ - pytest -v -s plugins_tests/test_oot_registration_online.py # it needs a clean process
621
+ - pytest -v -s plugins_tests/test_oot_registration_offline.py # it needs a clean process
622
+ - pytest -v -s plugins_tests/lora_resolvers # unit tests for in-tree lora resolver plugins
623
624
#------------------------------------------------------------ mi250 · v1 -------------------------------------------------------------#
625
@@ -1187,39 +1187,39 @@ steps:
1187
1188
#-------------------------------------------------------- mi300 · entrypoints --------------------------------------------------------#
1189
1190
-- label: Entrypoints Integration (API Server 2) # TBD
1190
+- label: Entrypoints Unit Tests # TBD
1191
timeout_in_minutes: 180
1192
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1193
agent_pool: mi300_1
1194
- optional: true
1194
fast_check: true
1196
- torch_nightly: true
1195
working_dir: "/vllm-workspace/tests"
1196
source_file_dependencies:
1199
- - vllm/
1200
- - tests/entrypoints/serve
1197
+ - vllm/entrypoints
1198
+ - tests/entrypoints/unit_tests
1199
+ - tests/entrypoints/weight_transfer
1200
+ - vllm/platforms/rocm.py
1201
commands:
1202
- - export VLLM_WORKER_MULTIPROC_METHOD=spawn
1203
- - pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc
1204
- - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/serve/dev/rpc
1202
+ - pytest -v -s entrypoints/unit_tests
1203
+ - pytest -v -s entrypoints/weight_transfer
1204
1206
-- label: Entrypoints Integration (API Server openai - Part 1) # TBD
1205
+- label: Entrypoints Integration (LLM) # TBD
1206
timeout_in_minutes: 180
1207
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1208
agent_pool: mi300_1
1209
+ optional: true
1210
fast_check: true
1211
torch_nightly: true
1212
- optional: true
1212
working_dir: "/vllm-workspace/tests"
1213
source_file_dependencies:
1214
- vllm/
1216
- - tests/entrypoints/openai
1217
- - tests/entrypoints/test_chat_utils
1215
+ - tests/entrypoints/llm
1216
commands:
1217
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1220
- - pytest -v -s entrypoints/openai/chat_completion --ignore=entrypoints/openai/chat_completion/test_oot_registration.py
1218
+ - pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py --ignore=entrypoints/llm/offline_mode
1219
+ - pytest -v -s entrypoints/llm/test_generate.py # it needs a clean process
1220
+ - pytest -v -s entrypoints/llm/offline_mode # Needs to avoid interference with other tests
1221
1222
-- label: Entrypoints Integration (API Server openai - Part 2) # TBD
1222
+- label: Entrypoints Integration (API Server) # TBD
1223
timeout_in_minutes: 180
1224
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1225
agent_pool: mi300_1
@@ -1229,24 +1229,19 @@ steps:
1229
working_dir: "/vllm-workspace/tests"
1230
source_file_dependencies:
1231
- vllm/
1232
- - tests/entrypoints/openai
1233
- - tests/entrypoints/test_chat_utils
1234
- - tests/entrypoints/generate
1235
- - tests/tool_use
1232
+ - tests/entrypoints/serve
1233
commands:
1234
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1238
- - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py
1239
- - pytest -v -s entrypoints/test_chat_utils.py
1240
- - pytest -v -s entrypoints/generate
1241
- - pytest -v -s tool_use
1235
+ - pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc
1236
+ - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/serve/dev/rpc
1237
1243
-- label: Entrypoints Integration (API Server openai - Part 3) # TBD
1238
+- label: Entrypoints Integration (API Server OpenAI - Part 1) # TBD
1239
timeout_in_minutes: 180
1240
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1241
agent_pool: mi300_1
1247
- optional: true
1242
fast_check: true
1243
torch_nightly: true
1244
+ optional: true
1245
working_dir: "/vllm-workspace/tests"
1246
source_file_dependencies:
1247
- vllm/
@@ -1254,54 +1249,61 @@ steps:
1249
- tests/entrypoints/test_chat_utils
1250
commands:
1251
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1257
- - pytest -v -s entrypoints/openai --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/correctness/ --ignore=entrypoints/openai/tool_parsers/ --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/test_multi_api_servers.py
1252
+ - pytest -v -s entrypoints/openai/ --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/correctness
1253
1259
-- label: Entrypoints Integration (Speech to Text) # TBD
1254
+- label: Entrypoints Integration (API Server OpenAI - Part 2) # TBD
1255
timeout_in_minutes: 180
1256
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1257
agent_pool: mi300_1
1258
fast_check: true
1259
torch_nightly: true
1260
+ optional: true
1261
working_dir: "/vllm-workspace/tests"
1262
source_file_dependencies:
1263
- vllm/
1268
- - tests/entrypoints/speech_to_text
1264
+ - tests/entrypoints/openai
1265
+ - tests/entrypoints/test_chat_utils
1266
commands:
1267
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1271
- - pytest -v -s entrypoints/speech_to_text
1268
+ - pytest -v -s entrypoints/openai/chat_completion
1269
+ - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py
1270
1273
-- label: Entrypoints Integration (Multimodal)
1271
+- label: Entrypoints Integration (API Server Generate) # TBD
1272
timeout_in_minutes: 180
1273
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1274
agent_pool: mi300_1
1275
+ optional: true
1276
fast_check: true
1277
torch_nightly: true
1278
working_dir: "/vllm-workspace/tests"
1279
source_file_dependencies:
1280
- vllm/
1282
- - tests/entrypoints/multimodal
1281
+ - tests/tool_use
1282
+ - tests/entrypoints/tool_parsers
1283
+ - tests/entrypoints/anthropic
1284
+ - tests/entrypoints/generate
1285
commands:
1286
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1285
- - pytest -v -s entrypoints/multimodal
1287
+ - pytest -v -s tool_use
1288
+ - pytest -v -s entrypoints/tool_parsers
1289
+ - pytest -v -s entrypoints/generate
1290
+ - pytest -v -s entrypoints/anthropic
1291
1287
-- label: Entrypoints Integration (LLM) # TBD
1292
+- label: Entrypoints Integration (Responses API) # TBD
1293
timeout_in_minutes: 180
1294
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1295
agent_pool: mi300_1
1291
- optional: true
1296
fast_check: true
1297
torch_nightly: true
1298
working_dir: "/vllm-workspace/tests"
1299
source_file_dependencies:
1300
- vllm/
1297
- - tests/entrypoints/llm
1301
+ - tests/entrypoints/openai/responses
1302
commands:
1303
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1300
- - pytest -v -s entrypoints/llm --ignore=entrypoints/llm/test_generate.py --ignore=entrypoints/llm/test_collective_rpc.py --ignore=entrypoints/llm/offline_mode
1301
- - pytest -v -s entrypoints/llm/test_generate.py # it needs a clean process
1302
- - pytest -v -s entrypoints/llm/offline_mode # Needs to avoid interference with other tests
1304
+ - pytest -v -s entrypoints/openai/responses
1305
1304
-- label: Entrypoints Integration (Pooling) # TBD
1306
+- label: Entrypoints Integration (Speech to Text) # TBD
1307
timeout_in_minutes: 180
1308
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1309
agent_pool: mi300_1
@@ -1310,12 +1312,12 @@ steps:
1312
working_dir: "/vllm-workspace/tests"
1313
source_file_dependencies:
1314
- vllm/
1313
- - tests/entrypoints/pooling
1315
+ - tests/entrypoints/speech_to_text
1316
commands:
1317
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1316
- - pytest -v -s entrypoints/pooling
1318
+ - pytest -v -s entrypoints/speech_to_text
1319
1318
-- label: Entrypoints Integration (Responses API) # TBD
1320
+- label: Entrypoints Integration (Multimodal)
1321
timeout_in_minutes: 180
1322
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1323
agent_pool: mi300_1
@@ -1324,24 +1326,24 @@ steps:
1326
working_dir: "/vllm-workspace/tests"
1327
source_file_dependencies:
1328
- vllm/
1327
- - tests/entrypoints/openai/responses
1329
+ - tests/entrypoints/multimodal
1330
commands:
1331
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
1330
- - pytest -v -s entrypoints/openai/responses
1332
+ - pytest -v -s entrypoints/multimodal
1333
1332
-- label: Entrypoints Unit Tests # TBD
1334
+- label: Entrypoints Integration (Pooling) # TBD
1335
timeout_in_minutes: 180
1336
mirror_hardwares: [amdexperimental, amdproduction, amdgfx942nightly, amdmi300]
1337
agent_pool: mi300_1
1338
fast_check: true
1339
+ torch_nightly: true
1340
working_dir: "/vllm-workspace/tests"
1341
source_file_dependencies:
1339
- - vllm/entrypoints
1340
- - tests/entrypoints/
1341
- - vllm/platforms/rocm.py
1342
+ - vllm/
1343
+ - tests/entrypoints/pooling
1344
commands:
1343
- - pytest -v -s entrypoints/openai/tool_parsers
1344
- - pytest -v -s entrypoints/ --ignore=entrypoints/llm --ignore=entrypoints/openai --ignore=entrypoints/serve --ignore=entrypoints/test_chat_utils.py --ignore=entrypoints/pooling --ignore=entrypoints/speech_to_text --ignore=entrypoints/pooling --ignore=entrypoints/multimodal --ignore=entrypoints/generate
1345
+ - export VLLM_WORKER_MULTIPROC_METHOD=spawn
1346
+ - pytest -v -s entrypoints/pooling
1347
1348
- label: OpenAI API correctness # TBD
1349
timeout_in_minutes: 180
@@ -1352,7 +1354,6 @@ steps:
1354
source_file_dependencies:
1355
- csrc/
1356
- vllm/entrypoints/openai/
1355
- - vllm/model_executor/models/whisper.py
1357
- vllm/model_executor/layers/
1358
- vllm/v1/attention/backends/
1359
- vllm/v1/attention/selector.py
@@ -2797,7 +2798,7 @@ steps:
2798
2799
#-------------------------------------------------------- mi355 · entrypoints --------------------------------------------------------#
2800
2800
-- label: Entrypoints Integration (API Server 2) # TBD
2801
+- label: Entrypoints Integration (API Server) # TBD
2802
timeout_in_minutes: 180
2803
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
2804
agent_pool: mi355_1
@@ -2813,7 +2814,7 @@ steps:
2814
- pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc
2815
- PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/serve/dev/rpc
2816
2816
-- label: Entrypoints Integration (API Server openai - Part 1) # TBD
2817
+- label: Entrypoints Integration (API Server OpenAI - Part 1) # TBD
2818
timeout_in_minutes: 180
2819
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
2820
agent_pool: mi355_1
@@ -2827,9 +2828,9 @@ steps:
2828
- tests/entrypoints/test_chat_utils
2829
commands:
2830
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
2830
- - pytest -v -s entrypoints/openai/chat_completion --ignore=entrypoints/openai/chat_completion/test_oot_registration.py
2831
+ - pytest -v -s entrypoints/openai --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/correctness
2832
2832
-- label: Entrypoints Integration (API Server openai - Part 2) # TBD
2833
+- label: Entrypoints Integration (API Server OpenAI - Part 2) # TBD
2834
timeout_in_minutes: 180
2835
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
2836
agent_pool: mi355_1
@@ -2841,29 +2842,31 @@ steps:
2842
- vllm/
2843
- tests/entrypoints/openai
2844
- tests/entrypoints/test_chat_utils
2844
- - tests/entrypoints/generate
2845
- - tests/tool_use
2845
commands:
2846
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
2847
+ - pytest -v -s entrypoints/openai/chat_completion
2848
- pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py
2849
- - pytest -v -s entrypoints/test_chat_utils.py
2850
- - pytest -v -s entrypoints/generate
2851
- - pytest -v -s tool_use
2849
2853
-- label: Entrypoints Integration (API Server openai - Part 3) # TBD
2850
+- label: Entrypoints Integration (API Server Generate) # TBD
2851
timeout_in_minutes: 180
2852
mirror_hardwares: [amdexperimental, amdproduction, amdgfx950nightly, amdmi355]
2853
agent_pool: mi355_1
2854
fast_check: true
2855
torch_nightly: true
2856
+ optional: true
2857
working_dir: "/vllm-workspace/tests"
2858
source_file_dependencies:
2859
- vllm/
2862
- - tests/entrypoints/openai
2863
- - tests/entrypoints/test_chat_utils
2860
+ - tests/tool_use
2861
+ - tests/entrypoints/tool_parsers
2862
+ - tests/entrypoints/anthropic
2863
+ - tests/entrypoints/generate
2864
commands:
2865
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
2866
- - pytest -v -s entrypoints/openai --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/correctness/ --ignore=entrypoints/openai/tool_parsers/ --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/test_multi_api_servers.py
2866
+ - pytest -v -s tool_use
2867
+ - pytest -v -s entrypoints/tool_parsers
2868
+ - pytest -v -s entrypoints/generate
2869
+ - pytest -v -s entrypoints/anthropic
2870
2871
- label: Entrypoints Integration (Speech to Text) # TBD
2872
timeout_in_minutes: 180
.buildkite/test_areas/entrypoints.yaml
+44
-44
@@ -8,10 +8,11 @@ steps:
8
working_dir: "/vllm-workspace/tests"
9
source_file_dependencies:
10
- vllm/entrypoints
11
- - tests/entrypoints/
11
+ - tests/entrypoints/unit_tests
12
+ - tests/entrypoints/weight_transfer
13
commands:
13
- - pytest -v -s entrypoints/openai/tool_parsers
14
- - pytest -v -s entrypoints/ --ignore=entrypoints/llm --ignore=entrypoints/openai --ignore=entrypoints/serve --ignore=entrypoints/test_chat_utils.py --ignore=entrypoints/pooling --ignore=entrypoints/speech_to_text --ignore=entrypoints/multimodal --ignore=entrypoints/generate
14
+ - pytest -v -s entrypoints/unit_tests
15
+ - pytest -v -s entrypoints/weight_transfer
16
17
- label: Entrypoints Integration (LLM)
18
key: entrypoints-integration-llm
@@ -31,50 +32,45 @@ steps:
32
depends_on:
33
- image-build-amd
34
34
-- label: Entrypoints Integration (API Server openai - Part 1)
35
- key: entrypoints-integration-api-server-openai-part-1
36
- timeout_in_minutes: 50
35
+- label: Entrypoints Integration (API Server)
36
+ key: entrypoints-integration-api-server
37
+ device: h200_35gb
38
+ timeout_in_minutes: 130
39
working_dir: "/vllm-workspace/tests"
40
source_file_dependencies:
41
- vllm/
40
- - tests/entrypoints/openai
41
- - tests/entrypoints/test_chat_utils
42
+ - tests/entrypoints/serve
43
commands:
44
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
44
- - pytest -v -s entrypoints/openai/chat_completion --ignore=entrypoints/openai/chat_completion/test_oot_registration.py
45
+ - pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc
46
+ - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/serve/dev/rpc
47
mirror:
48
amd:
49
device: mi300_1
48
- timeout_in_minutes: 80
50
depends_on:
51
- image-build-amd
52
52
-- label: Entrypoints Integration (API Server openai - Part 2)
53
- key: entrypoints-integration-api-server-openai-part-2
53
+- label: Entrypoints Integration (API Server OpenAI - Part 1)
54
+ key: entrypoints-integration-api-server-openai-part-1
55
timeout_in_minutes: 50
56
working_dir: "/vllm-workspace/tests"
57
source_file_dependencies:
58
- vllm/
59
- tests/entrypoints/openai
60
- tests/entrypoints/test_chat_utils
60
- - tests/entrypoints/generate
61
- - tests/tool_use
61
commands:
63
- - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py
64
- - pytest -v -s entrypoints/test_chat_utils.py
65
- - pytest -v -s entrypoints/generate
66
- - pytest -v -s tool_use
62
+ - export VLLM_WORKER_MULTIPROC_METHOD=spawn
63
+ - pytest -v -s entrypoints/openai --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/correctness
64
mirror:
65
amd:
66
device: mi300_1
70
- timeout_in_minutes: 60
67
+ timeout_in_minutes: 80
68
depends_on:
69
- image-build-amd
70
74
-- label: Entrypoints Integration (API Server openai - Part 3)
75
- key: entrypoints-integration-api-server-openai-part-3
71
+- label: Entrypoints Integration (API Server OpenAI - Part 2)
72
+ key: entrypoints-integration-api-server-openai-part-2
73
timeout_in_minutes: 50
77
- device: h200_18gb
74
working_dir: "/vllm-workspace/tests"
75
source_file_dependencies:
76
- vllm/
@@ -82,32 +78,47 @@ steps:
78
- tests/entrypoints/test_chat_utils
79
commands:
80
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
85
- - pytest -v -s entrypoints/openai --ignore=entrypoints/openai/chat_completion --ignore=entrypoints/openai/completion --ignore=entrypoints/openai/correctness/ --ignore=entrypoints/openai/tool_parsers/ --ignore=entrypoints/openai/responses --ignore=entrypoints/openai/test_multi_api_servers.py
81
+ - pytest -v -s entrypoints/openai/chat_completion
82
+ - pytest -v -s entrypoints/openai/completion --ignore=entrypoints/openai/completion/test_tensorizer_entrypoint.py
83
mirror:
84
amd:
88
- device: mi325_1
89
- timeout_in_minutes: 60
85
+ device: mi300_1
86
+ timeout_in_minutes: 80
87
depends_on:
88
- image-build-amd
89
93
-- label: Entrypoints Integration (API Server 2)
94
- device: h200_35gb
95
- key: entrypoints-integration-api-server-2
96
- timeout_in_minutes: 130
90
+- label: Entrypoints Integration (API Server Generate)
91
+ key: entrypoints-integration-api-server-generate
92
+ timeout_in_minutes: 50
93
working_dir: "/vllm-workspace/tests"
94
source_file_dependencies:
95
- vllm/
100
- - tests/entrypoints/serve
96
+ - tests/tool_use
97
+ - tests/entrypoints/tool_parsers
98
+ - tests/entrypoints/anthropic
99
+ - tests/entrypoints/generate
100
commands:
102
- - export VLLM_WORKER_MULTIPROC_METHOD=spawn
103
- - pytest -v -s entrypoints/serve --ignore=entrypoints/serve/dev/rpc
104
- - PYTHONPATH=/vllm-workspace pytest -v -s entrypoints/serve/dev/rpc
101
+ - pytest -v -s tool_use
102
+ - pytest -v -s entrypoints/tool_parsers
103
+ - pytest -v -s entrypoints/generate
104
+ - pytest -v -s entrypoints/anthropic
105
mirror:
106
amd:
107
device: mi300_1
108
+ timeout_in_minutes: 60
109
depends_on:
110
- image-build-amd
111
112
+- label: Entrypoints Integration (Responses API)
113
+ key: entrypoints-integration-responses-api
114
+ timeout_in_minutes: 50
115
+ working_dir: "/vllm-workspace/tests"
116
+ source_file_dependencies:
117
+ - vllm/
118
+ - tests/entrypoints/openai/responses
119
+ commands:
120
+ - pytest -v -s entrypoints/openai/responses
121
+
122
- label: Entrypoints Integration (Speech to Text)
123
device: h200_35gb
124
key: entrypoints-integration-speech_to_text
@@ -143,16 +154,6 @@ steps:
154
- export VLLM_WORKER_MULTIPROC_METHOD=spawn
155
- pytest -v -s entrypoints/pooling
156
146
-- label: Entrypoints Integration (Responses API)
147
- key: entrypoints-integration-responses-api
148
- timeout_in_minutes: 50
149
- working_dir: "/vllm-workspace/tests"
150
- source_file_dependencies:
151
- - vllm/
152
- - tests/entrypoints/openai/responses
153
- commands:
154
- - pytest -v -s entrypoints/openai/responses
155
-
157
- label: OpenAI API Correctness
158
key: openai-api-correctness
159
timeout_in_minutes: 30
@@ -170,7 +171,6 @@ steps:
171
source_file_dependencies:
172
- csrc/
173
- vllm/entrypoints/openai/
173
- - vllm/model_executor/models/whisper.py
174
- vllm/model_executor/layers/
175
- vllm/v1/attention/backends/
176
- vllm/v1/attention/selector.py
.buildkite/test_areas/plugins.yaml
+3
-3
@@ -37,6 +37,6 @@ steps:
37
- pytest -v -s plugins_tests/test_scheduler_plugins.py
38
- pip install -e ./plugins/vllm_add_dummy_model
39
- pytest -v -s distributed/test_distributed_oot.py
40
- - pytest -v -s entrypoints/openai/chat_completion/test_oot_registration.py # it needs a clean process
41
- - pytest -v -s models/test_oot_registration.py # it needs a clean process
42
- - pytest -v -s plugins/lora_resolvers # unit tests for in-tree lora resolver plugins
40
+ - pytest -v -s plugins_tests/test_oot_registration_online.py # it needs a clean process
41
+ - pytest -v -s plugins_tests/test_oot_registration_offline.py # it needs a clean process
42
+ - pytest -v -s plugins_tests/lora_resolvers # unit tests for in-tree lora resolver plugins
tests/distributed/test_distributed_oot.py
+1
-1
@@ -1,7 +1,7 @@
1
# SPDX-License-Identifier: Apache-2.0
2
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
3
4
-from tests.entrypoints.openai.chat_completion.test_oot_registration import (
4
+from tests.plugins_tests.test_oot_registration_online import (
5
run_and_test_dummy_opt_api_server,
6
)
7
tests/entrypoints/tool_parsers/__init__.py
renamed
tests/entrypoints/tool_parsers/test_granite4_tool_parser.py
renamed
+1
-1
@@ -5,7 +5,7 @@ import json
5
import openai
6
import pytest
7
8
-from ....utils import RemoteOpenAIServer
8
+from tests.utils import RemoteOpenAIServer
9
10
MODEL = "ibm-granite/granite-4.0-h-tiny"
11
tests/entrypoints/tool_parsers/test_hermes_tool_parser.py
renamed
+1
-2
@@ -9,12 +9,11 @@ import pytest_asyncio
9
from huggingface_hub import snapshot_download
10
from typing_extensions import TypedDict
11
12
+from tests.utils import RemoteOpenAIServer
13
from vllm.tool_parsers.abstract_tool_parser import ToolParser
14
from vllm.tool_parsers.granite4_tool_parser import Granite4ToolParser
15
from vllm.tool_parsers.hermes_tool_parser import Hermes2ProToolParser
16
16
-from ....utils import RemoteOpenAIServer
17
-
17
LORA_MODEL = "minpeter/LoRA-Llama-3.2-1B-tool-vllm-ci"
18
19
TOOLS = [
tests/entrypoints/tool_parsers/test_openai_tool_parser.py
renamed
+1
-1
@@ -9,7 +9,7 @@ import pytest
9
import pytest_asyncio
10
from rapidfuzz import fuzz
11
12
-from ....utils import RemoteOpenAIServer
12
+from tests.utils import RemoteOpenAIServer
13
14
MODEL_NAME = "openai/gpt-oss-20b"
15
tests/entrypoints/unit_tests/__init__.py
renamed
tests/entrypoints/unit_tests/test_api_server_process_manager.py
renamed
tests/entrypoints/unit_tests/test_chat_utils.py
renamed
tests/entrypoints/unit_tests/test_context.py
renamed
tests/entrypoints/unit_tests/test_grpc_health.py
renamed
tests/entrypoints/unit_tests/test_launch_cli.py
renamed
tests/plugins_tests/__init__.py
tests/plugins_tests/lora_resolvers/__init__.py
tests/plugins_tests/lora_resolvers/test_filesystem_resolver.py
renamed
tests/plugins_tests/lora_resolvers/test_hf_hub_resolver.py
renamed
tests/plugins_tests/test_oot_registration_offline.py
renamed
+1
-2
@@ -3,12 +3,11 @@
3
4
import pytest
5
6
+from tests.utils import create_new_process_for_each_test
7
from vllm import LLM, SamplingParams
8
from vllm.assets.image import ImageAsset
9
from vllm.multimodal.image import convert_image_mode
10
10
-from ..utils import create_new_process_for_each_test
11
-
11
12
@create_new_process_for_each_test()
13
def test_plugin(