@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 6bd19551b0
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 16, 18:15 UTC

.[Build/CI] Enabling passing AMD tests. (#7610)

Alexei-V-Ivanov-AMD committed Aug 16, 2024 at 22:25 UTC 6bd19551b0fa8083e6ed1cb35f5375499d855fce
1 file changed +6 -7
.buildkite/test-pipeline.yaml
+6 -7
@@ -56,7 +56,7 @@ steps:
56 - pytest -v -s worker # Worker
57
58 - label: Basic Correctness Test # 30min
59 - mirror_hardwares: [amd]
59 + #mirror_hardwares: [amd]
60 fast_check: true
61 source_file_dependencies:
62 - vllm/
@@ -81,7 +81,7 @@ steps:
81 - label: Entrypoints Test # 20min
82 working_dir: "/vllm-workspace/tests"
83 fast_check: true
84 - mirror_hardwares: [amd]
84 + #mirror_hardwares: [amd]
85 source_file_dependencies:
86 - vllm/
87 commands:
@@ -141,7 +141,7 @@ steps:
141
142 - label: Examples Test # 12min
143 working_dir: "/vllm-workspace/examples"
144 - mirror_hardwares: [amd]
144 + #mirror_hardwares: [amd]
145 source_file_dependencies:
146 - vllm/entrypoints
147 - examples/
@@ -173,14 +173,14 @@ steps:
173
174
175 - label: Vision Language Models Test # 42min
176 - mirror_hardwares: [amd]
176 + #mirror_hardwares: [amd]
177 source_file_dependencies:
178 - vllm/
179 commands:
180 - pytest -v -s models -m vlm
181
182 - label: Prefix Caching Test # 7min
183 - mirror_hardwares: [amd]
183 + #mirror_hardwares: [amd]
184 source_file_dependencies:
185 - vllm/
186 - tests/prefix_caching
@@ -263,7 +263,6 @@ steps:
263 - export VLLM_WORKER_MULTIPROC_METHOD=spawn
264 - bash ./run-tests.sh -c configs/models-small.txt -t 1
265
266 -
266 ##### 1 GPU test #####
267 ##### multi gpus test #####
268
@@ -292,7 +291,7 @@ steps:
291 - VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py
292
293 - label: Distributed Tests (2 GPUs) # 28min
295 - mirror_hardwares: [amd]
294 + #mirror_hardwares: [amd]
295 working_dir: "/vllm-workspace/tests"
296 num_gpus: 2
297 source_file_dependencies: