biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[CI/Build] Split up decoder-only LM tests (#9488)

Browse Source 
Co-authored-by: Nick Hill <nickhill@us.ibm.com>
This commit is contained in:
Cyrus Leung
2024-10-21 12:27:50 +08:00
committed by GitHub
parent 855e0e6f97
commit 696b01af8f
3 changed files with 18 additions and 57 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
tests/models/decoder_only/language/test_big_models.py
View File

@@ -21,10 +21,14 @@ MODELS = [
]
if not current_platform.is_cpu():
# MiniCPM requires fused_moe which is not supported by CPU
MODELS.append("openbmb/MiniCPM3-4B")
MODELS += [
# fused_moe which not supported on CPU
"openbmb/MiniCPM3-4B",
# Head size isn't supported on CPU
"h2oai/h2o-danube3-4b-base",
]
#TODO: remove this after CPU float16 support ready
# TODO: remove this after CPU float16 support ready
target_dtype = "float" if current_platform.is_cpu() else "half"