xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-01-26 05:04:28 +08:00
Code Issues Packages Projects Releases Wiki Activity

[ROCm][CI] Fix Weight Loading With Multiple GPU Tests on ROCm (#28984)

Browse Source 
Signed-off-by: Micah Williamson <micah.williamson@amd.com>
This commit is contained in:
Micah Williamson 2025-11-19 15:31:33 -06:00 committed by GitHub
parent 88f5b19f0b
commit 22e44ad589
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 8 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
.buildkite/test-amd.yaml
View File

@ -1323,7 +1323,7 @@ steps:
- vllm/
- tests/weight_loading
commands:
- bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models.txt
- bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-amd.txt
- label: Weight Loading Multiple GPU Test - Large Models # optional
mirror_hardwares: [amdexperimental]
@ -1331,13 +1331,12 @@ steps:
# grade: Blocking
working_dir: "/vllm-workspace/tests"
num_gpus: 2
gpu: a100
optional: true
source_file_dependencies:
- vllm/
- tests/weight_loading
commands:
- bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-large.txt
- bash weight_loading/run_model_weight_loading_test.sh -c weight_loading/models-large-amd.txt
- label: NixlConnector PD accuracy tests (Distributed) # 30min
mirror_hardwares: [amdexperimental]

3
tests/weight_loading/models-amd.txt Normal file
View File

@ -0,0 +1,3 @@
fp8, amd/Meta-Llama-3.1-8B-Instruct-FP8-KV, main
None, amd/Llama-3.2-1B-Instruct-FP8-KV, main
fp8, amd/Mixtral-8x7B-Instruct-v0.1-FP8-KV, main

3
tests/weight_loading/models-large-amd.txt Normal file
View File

@ -0,0 +1,3 @@
fp8, amd/Meta-Llama-3.1-70B-Instruct-FP8-KV, main
None, microsoft/phi-4, main
fp8, amd/Mixtral-8x22B-Instruct-v0.1-FP8-KV, main