[1/n][CI] Load models in CI from S3 instead of HF (#13205)

Signed-off-by: <>
Co-authored-by: EC2 Default User <ec2-user@ip-172-31-20-117.us-west-2.compute.internal>
This commit is contained in:
Kevin H. Luu
2025-02-19 07:34:59 +00:00
committed by GitHub
co-authored by EC2 Default User
parent fd84857f64
commit d5d214ac7f
43 changed files with 225 additions and 76 deletions
@@ -27,6 +27,8 @@ from vllm.model_executor.layers.quantization import (QuantizationConfig,
from vllm.platforms import current_platform
from vllm.utils import PlaceholderModule
logger = init_logger(__name__)
try:
from runai_model_streamer import SafetensorsStreamer
except (ImportError, OSError):
@@ -37,8 +39,6 @@ except (ImportError, OSError):
SafetensorsStreamer = runai_model_streamer.placeholder_attr(
"SafetensorsStreamer")
logger = init_logger(__name__)
# use system-level temp directory for file locks, so that multiple users
# can share the same lock without error.
# lock files in the temp directory will be automatically deleted when the