@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 40c27a7cbb
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, 19:30 UTC

[Build] Temporarily Disable Kernels and LoRA tests (#6961)

Simon Mo committed Jul 30, 2024 at 14:59 UTC 40c27a7cbb496ede63da9d636c07a1f315fd36e1
1 file changed +20 -20
.buildkite/test-pipeline.yaml
+20 -20
@@ -155,12 +155,12 @@ steps:
155 - pytest -v -s test_inputs.py
156 - pytest -v -s multimodal
157
158 -- label: Kernels Test %N
159 - #mirror_hardwares: [amd]
160 - commands:
161 - - pip install https://github.com/flashinfer-ai/flashinfer/releases/download/v0.0.8/flashinfer-0.0.8+cu121torch2.3-cp310-cp310-linux_x86_64.whl
162 - - pytest -v -s kernels --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
163 - parallelism: 4
158 +# - label: Kernels Test %N
159 +# #mirror_hardwares: [amd]
160 +# commands:
161 +# - pip install https://github.com/flashinfer-ai/flashinfer/releases/download/v0.0.8/flashinfer-0.0.8+cu121torch2.3-cp310-cp310-linux_x86_64.whl
162 +# - pytest -v -s kernels --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT
163 +# parallelism: 4
164
165 - label: Models Test
166 #mirror_hardwares: [amd]
@@ -202,20 +202,20 @@ steps:
202 - export VLLM_ATTENTION_BACKEND=XFORMERS
203 - pytest -v -s spec_decode
204
205 -- label: LoRA Test %N
206 - #mirror_hardwares: [amd]
207 - command: pytest -v -s lora --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --ignore=lora/test_long_context.py
208 - parallelism: 4
209 -
210 -- label: LoRA Long Context (Distributed)
211 - #mirror_hardwares: [amd]
212 - num_gpus: 4
213 - # This test runs llama 13B, so it is required to run on 4 GPUs.
214 - commands:
215 - # FIXIT: find out which code initialize cuda before running the test
216 - # before the fix, we need to use spawn to test it
217 - - export VLLM_WORKER_MULTIPROC_METHOD=spawn
218 - - pytest -v -s -x lora/test_long_context.py
205 +# - label: LoRA Test %N
206 +# #mirror_hardwares: [amd]
207 +# command: pytest -v -s lora --shard-id=$$BUILDKITE_PARALLEL_JOB --num-shards=$$BUILDKITE_PARALLEL_JOB_COUNT --ignore=lora/test_long_context.py
208 +# parallelism: 4
209 +
210 +# - label: LoRA Long Context (Distributed)
211 +# #mirror_hardwares: [amd]
212 +# num_gpus: 4
213 +# # This test runs llama 13B, so it is required to run on 4 GPUs.
214 +# commands:
215 +# # FIXIT: find out which code initialize cuda before running the test
216 +# # before the fix, we need to use spawn to test it
217 +# - export VLLM_WORKER_MULTIPROC_METHOD=spawn
218 +# - pytest -v -s -x lora/test_long_context.py
219
220 - label: Tensorizer Test
221 #mirror_hardwares: [amd]