Mirrors
/
hf_text-generation-inference
mirror of https://github.com/huggingface/text-generation-inference.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

fix: only run test when cuda is available

This commit is contained in:
David Holtz 2024-10-09 18:43:53 +00:00
parent a8108bc0da
commit 1ddde382bd
1 changed files with 12 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
server/tests/utils/test_kv_cache.py
View File

@ -1,14 +1,12 @@
import torch import torch
import pytest import pytest
from text_generation_server.models.globals import ATTENTION, BLOCK_SIZE from text_generation_server.models.globals import ATTENTION, BLOCK_SIZE
from text_generation_server.layers.attention import KVCache
from text_generation_server.utils.import_utils import SYSTEM from text_generation_server.utils.import_utils import SYSTEM
def test_kvcache_memory(): # only include this import when CUDA is available
if SYSTEM == "cuda": if SYSTEM == "cuda":
kvcache_memory() from text_generation_server.layers.attention import KVCache
else:
pytest.skip("Test only runs on CUDA")
def kvcache_memory(): def kvcache_memory():
num_blocks = 8188 num_blocks = 8188
@ -40,5 +38,12 @@ def kvcache_memory():
assert kv_cache_memory_mb < 1025 assert kv_cache_memory_mb < 1025
# only include this test when CUDA is available
if SYSTEM == "cuda":
def test_kvcache_memory():
kvcache_memory()
if __name__ == "__main__": if __name__ == "__main__":
test_kvcache_memory() test_kvcache_memory()