Skip to content

Commit ddfc749

Browse files
zhulinJulia24zhulinJulia24
authored
[ci] update daily testcase (#3944)
* update * update * update * Update evaluate.yml * update * update * update * update * update * update * update * update * update * update * update * update datasets version * update * Update daily_ete_test_3090.yml * Update Docker image to use latest-cu12 version * update * Update benchmark.yml * update * update * update * update * update image to cu128 * updaet
1 parent a7d7a58 commit ddfc749

21 files changed

+600
-560
lines changed

‎.github/scripts/eval_regression_base_models.py‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -62,6 +62,7 @@
6262
pytorch_qwen2_5_72b_model = deepcopy(lmdeploy_qwen2_5_72b_model)
6363
pytorch_qwen2_7b_model = deepcopy(lmdeploy_qwen2_7b_model)
6464
pytorch_yi_1_5_9b_model = deepcopy(lmdeploy_yi_1_5_9b_model)
65+
pytorch_deepseek_v2_model['engine_config']['cache_max_entry_count'] = 0.6
6566

6667
lmdeploy_glm4_9b_model_native = deepcopy(lmdeploy_glm4_9b_model)
6768
lmdeploy_deepseek_7b_base_model_native = deepcopy(lmdeploy_deepseek_7b_base_model)

‎.github/scripts/eval_regression_chat_models.py‎

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -91,6 +91,7 @@
9191
pytorch_qwen2_5_32b_instruct_model = deepcopy(lmdeploy_qwen2_5_32b_instruct_model)
9292
pytorch_qwen2_7b_instruct_model = deepcopy(lmdeploy_qwen2_7b_instruct_model)
9393
pytorch_yi_1_5_34b_chat_model = deepcopy(lmdeploy_yi_1_5_34b_chat_model)
94+
pytorch_deepseek_v2_5_1210_model['engine_config']['cache_max_entry_count'] = 0.6
9495

9596
lmdeploy_glm4_9b_chat_model_native = deepcopy(lmdeploy_glm4_9b_chat_model)
9697
lmdeploy_deepseek_r1_distill_qwen_32b_model_native = deepcopy(lmdeploy_deepseek_r1_distill_qwen_32b_model)

‎.github/workflows/benchmark.yml‎

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -78,13 +78,14 @@ jobs:
7878
benchmark_type: ${{fromJSON(github.event.inputs.benchmark_type)}}
7979
timeout-minutes: 480
8080
container:
81-
image: openmmlab/lmdeploy:latest-cu11
81+
image: openmmlab/lmdeploy:latest-cu12.8
8282
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
8383
volumes:
8484
- /nvme/github-actions/pip-cache:/root/.cache/pip
8585
- /nvme/github-actions/packages:/root/packages
8686
- /nvme/qa_test_models:/nvme/qa_test_models
8787
- /mnt/shared:/mnt/shared
88+
- /mnt/bigdisk:/mnt/bigdisk
8889
- /usr/share/zoneinfo/Asia/Shanghai:/etc/localtime:ro
8990
steps:
9091
- name: Clone repository
@@ -143,7 +144,7 @@ jobs:
143144
timeout-minutes: 5
144145
runs-on: [self-hosted, linux-a100]
145146
container:
146-
image: openmmlab/lmdeploy:latest-cu11
147+
image: openmmlab/lmdeploy:latest-cu12.8
147148
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
148149
volumes:
149150
- /nvme/qa_test_models:/nvme/qa_test_models

‎.github/workflows/daily_ete_test.yml‎

Lines changed: 18 additions & 12 deletions
Original file line numberDiff line numberDiff line change
@@ -39,7 +39,7 @@ on:
3939
type: string
4040
default: "['quant', 'tools','restful','pipeline','benchmark','evaluation']"
4141
schedule:
42-
- cron: '00 16 * * 0-4'
42+
- cron: '00 14 * * 0-4'
4343

4444
env:
4545
HOST_PIP_CACHE_DIR: /nvme/github-actions/pip-cache
@@ -96,7 +96,7 @@ jobs:
9696
runs-on: [self-hosted, linux-a100]
9797
timeout-minutes: 50
9898
container:
99-
image: openmmlab/lmdeploy:latest-cu11
99+
image: openmmlab/lmdeploy:latest-cu12.8
100100
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
101101
volumes:
102102
- /nvme/qa_test_models:/nvme/qa_test_models
@@ -136,7 +136,7 @@ jobs:
136136
MODELSCOPE_CACHE: /nvme/qa_test_models/modelscope_hub
137137
MODELSCOPE_MODULES_CACHE: /nvme/qa_test_models/modelscope_modules
138138
container:
139-
image: openmmlab/lmdeploy:latest-cu11
139+
image: openmmlab/lmdeploy:latest-cu12.8
140140
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
141141
volumes:
142142
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -152,13 +152,15 @@ jobs:
152152
cp -r ${{env.TEST_CODE_PATH}}/. .
153153
- name: Install lmdeploy - dependency
154154
run: |
155-
python3 -m pip install -r ${{env.OFFLINE_REQUIREMENTS}}
155+
python3 -m pip install auto_gptq matplotlib attrdict
156+
python3 -m pip install -r requirements/lite.txt
156157
- name: Install lmdeploy
157158
run: |
158159
python3 -m pip uninstall lmdeploy -y && python3 -m pip install lmdeploy-*.whl --no-deps
159160
python3 -m pip install -r requirements/test.txt
160-
python3 -m pip install transformers==4.53.1
161+
rm -rf ${{env.DEEPSEEK_VL}}/build
161162
pip install ${{env.DEEPSEEK_VL}} --no-deps
163+
python3 -m pip install transformers==4.53.1 datasets==3.6.0 timm
162164
- name: Check env
163165
run: |
164166
python3 -m pip list
@@ -217,7 +219,7 @@ jobs:
217219
MODELSCOPE_CACHE: /nvme/qa_test_models/modelscope_hub
218220
MODELSCOPE_MODULES_CACHE: /nvme/qa_test_models/modelscope_modules
219221
container:
220-
image: openmmlab/lmdeploy:latest-cu11
222+
image: openmmlab/lmdeploy:latest-cu12.8
221223
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
222224
volumes:
223225
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -322,7 +324,7 @@ jobs:
322324
model: Intern-S1
323325
timeout-minutes: 60
324326
container:
325-
image: openmmlab/lmdeploy:latest-cu11
327+
image: openmmlab/lmdeploy:latest-cu12.8
326328
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
327329
volumes:
328330
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -341,6 +343,7 @@ jobs:
341343
run: |
342344
python3 -m pip uninstall lmdeploy -y && python3 -m pip install lmdeploy-*.whl --no-deps
343345
python3 -m pip install -r requirements/test.txt
346+
rm -rf ${{env.DEEPSEEK_VL}}/build
344347
pip install ${{env.DEEPSEEK_VL}} --no-deps
345348
- name: Check env
346349
run: |
@@ -405,7 +408,7 @@ jobs:
405408
needs: test_quantization
406409
timeout-minutes: 120
407410
container:
408-
image: openmmlab/lmdeploy:latest-cu11
411+
image: openmmlab/lmdeploy:latest-cu12.8
409412
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
410413
volumes:
411414
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -424,6 +427,7 @@ jobs:
424427
run: |
425428
python3 -m pip uninstall lmdeploy -y && python3 -m pip install lmdeploy-*.whl --no-deps
426429
python3 -m pip install -r requirements/test.txt
430+
rm -rf ${{env.DEEPSEEK_VL}}/build
427431
pip install ${{env.DEEPSEEK_VL}} --no-deps
428432
- name: Check env
429433
run: |
@@ -461,7 +465,7 @@ jobs:
461465
needs: test_quantization
462466
timeout-minutes: 120
463467
container:
464-
image: openmmlab/lmdeploy:latest-cu11
468+
image: openmmlab/lmdeploy:latest-cu12.8
465469
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
466470
volumes:
467471
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -480,6 +484,7 @@ jobs:
480484
run: |
481485
python3 -m pip uninstall lmdeploy -y && python3 -m pip install lmdeploy-*.whl --no-deps
482486
python3 -m pip install -r requirements/test.txt
487+
rm -rf ${{env.DEEPSEEK_VL}}/build
483488
pip install ${{env.DEEPSEEK_VL}} --no-deps
484489
- name: Check env
485490
run: |
@@ -515,7 +520,7 @@ jobs:
515520
matrix:
516521
evaluate_type: ['chat', 'base']
517522
container:
518-
image: openmmlab/lmdeploy:latest-cu11
523+
image: openmmlab/lmdeploy:latest-cu12.8
519524
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
520525
volumes:
521526
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -539,6 +544,7 @@ jobs:
539544
run: |
540545
python3 -m pip uninstall lmdeploy -y && python3 -m pip install lmdeploy-*.whl --no-deps
541546
python3 -m pip install -r requirements/test.txt
547+
rm -rf ${{env.DEEPSEEK_VL}}/build
542548
pip install ${{env.DEEPSEEK_VL}} --no-deps
543549
- name: Install opencompass
544550
run: |
@@ -588,7 +594,7 @@ jobs:
588594
timeout-minutes: 5
589595
runs-on: [self-hosted, linux-a100]
590596
container:
591-
image: openmmlab/lmdeploy:latest-cu11
597+
image: openmmlab/lmdeploy:latest-cu12.8
592598
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
593599
volumes:
594600
- /nvme/qa_test_models:/nvme/qa_test_models
@@ -613,7 +619,7 @@ jobs:
613619
needs: [test_tools, test_restful, test_pipeline, test_benchmark]
614620
timeout-minutes: 5
615621
container:
616-
image: openmmlab/lmdeploy:latest-cu11
622+
image: openmmlab/lmdeploy:latest-cu12.8
617623
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
618624
volumes:
619625
- /nvme/github-actions/pip-cache:/root/.cache/pip

‎.github/workflows/daily_ete_test_3090.yml‎

Lines changed: 11 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -39,7 +39,7 @@ on:
3939
type: string
4040
default: "['quant', 'tools']"
4141
schedule:
42-
- cron: '00 16 * * 0-4'
42+
- cron: '00 14 * * 0-4'
4343

4444
env:
4545
HOST_PIP_CACHE_DIR: /nvme/github-actions/pip-cache
@@ -49,7 +49,7 @@ env:
4949
REPORT_DIR: /nvme/qa_test_models/test-reports/${{ github.run_id }}
5050
COV_PARAM: --cov /opt/py3/lib/python3.10/site-packages/lmdeploy
5151
FAIL_CONFIG: ${{ github.event_name == 'schedule' && github.run_attempt != 1 && '--lf --lfnf none' || '--lf'}}
52-
TEST_CODE_PATH: /nvme/qa_test_models/test_pkg/lmdeploy
52+
TEST_CODE_PATH: /nvme/qa_test_models/test_pkg/lmdeploy/${{ github.run_id }}
5353
OFFLINE_CODE_PATH: /nvme/qa_test_models/offline_pkg/lmdeploy
5454
OFFLINE_REQUIREMENTS: /nvme/qa_test_models/offline_pkg/requirements.txt
5555

@@ -95,7 +95,7 @@ jobs:
9595
runs-on: [self-hosted, 3090-r1]
9696
timeout-minutes: 50
9797
container:
98-
image: openmmlab/lmdeploy:latest
98+
image: openmmlab/lmdeploy:latest-cu12
9999
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
100100
volumes:
101101
- /nvme/qa_test_models:/nvme/qa_test_models
@@ -136,7 +136,7 @@ jobs:
136136
MODELSCOPE_CACHE: /nvme/qa_test_models/modelscope_hub
137137
MODELSCOPE_MODULES_CACHE: /nvme/qa_test_models/modelscope_modules
138138
container:
139-
image: openmmlab/lmdeploy:latest
139+
image: openmmlab/lmdeploy:latest-cu12
140140
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
141141
volumes:
142142
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -149,11 +149,12 @@ jobs:
149149
cp -r ${{env.TEST_CODE_PATH}}/. .
150150
- name: Install lmdeploy - dependency
151151
run: |
152-
python3 -m pip install -r ${{env.OFFLINE_REQUIREMENTS}}
152+
python3 -m pip install auto_gptq matplotlib
153+
python3 -m pip install -r requirements/lite.txt
153154
- name: Install lmdeploy
154155
run: |
155156
python3 -m pip install lmdeploy-*.whl --no-deps
156-
python3 -m pip install transformers==4.53.1
157+
python3 -m pip install transformers==4.53.1 datasets==3.6.0 timm
157158
python3 -m pip install -r requirements/test.txt
158159
- name: Check env
159160
run: |
@@ -209,7 +210,7 @@ jobs:
209210
MODELSCOPE_CACHE: /nvme/qa_test_models/modelscope_hub
210211
MODELSCOPE_MODULES_CACHE: /nvme/qa_test_models/modelscope_modules
211212
container:
212-
image: openmmlab/lmdeploy:latest
213+
image: openmmlab/lmdeploy:latest-cu12
213214
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
214215
volumes:
215216
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -274,7 +275,7 @@ jobs:
274275
backend: ${{ fromJSON(inputs.backend || '["turbomind", "pytorch"]')}}
275276
timeout-minutes: 60
276277
container:
277-
image: openmmlab/lmdeploy:latest
278+
image: openmmlab/lmdeploy:latest-cu12
278279
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
279280
volumes:
280281
- /nvme/github-actions/pip-cache:/root/.cache/pip
@@ -315,7 +316,7 @@ jobs:
315316
- name: Test lmdeploy - restful api
316317
timeout-minutes: 75
317318
run: |
318-
pytest autotest/interface/restful/test_restful_chat_func.py -n 20 -m 'not not_${{matrix.backend}}' --alluredir=${{env.REPORT_DIR}}/interface-${{matrix.backend}} ${{env.COV_PARAM}} || true
319+
pytest autotest/interface/restful/test_restful_chat_func.py -n 20 -m 'not not_${{matrix.backend}} and not interns1 and not internlm2_5' --alluredir=${{env.REPORT_DIR}}/interface-${{matrix.backend}} ${{env.COV_PARAM}} || true
319320
mv .coverage ${{env.REPORT_DIR}}/.coverage.$(date +'%Y%m%d%H%M%S')
320321
- name: Kill api server
321322
if: always()
@@ -358,7 +359,7 @@ jobs:
358359
needs: [test_tools, test_restful]
359360
timeout-minutes: 5
360361
container:
361-
image: openmmlab/lmdeploy:latest
362+
image: openmmlab/lmdeploy:latest-cu12
362363
options: "--gpus=all --ipc=host --user root -e PIP_CACHE_DIR=/root/.cache/pip -e NVIDIA_DISABLE_REQUIRE=1 --pull never"
363364
volumes:
364365
- /nvme/github-actions/pip-cache:/root/.cache/pip

0 commit comments

Comments
 (0)