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

[V1] V1 Enablement Oracle (#13726)

Browse Source 
Signed-off-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
Co-authored-by: rshaw@neuralmagic.com <rshaw@neuralmagic.com>
Co-authored-by: Nicolò Lucchesi <nlucches@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
This commit is contained in:
Robert Shaw
2025-03-15 01:02:20 -04:00
committed by GitHub
parent 8c0d15d5c5
commit d4d93db2c5
96 changed files with 1537 additions and 512 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
tests/weight_loading/test_weight_loading.py
View File

@@ -15,6 +15,9 @@ QUANTIZATION = os.environ.get("QUANTIZATION", "gptq_marlin")
MIN_CAPABILITY = os.environ.get("MIN_CAPABILITY", "80")
@pytest.mark.skipif(
MODEL_NAME == "casperhansen/deepseek-coder-v2-instruct-awq",
reason="OOM in the CI")
@pytest.mark.skipif(
not current_platform.has_device_capability(int(MIN_CAPABILITY)),
reason="Current system does not have minimum capability.")
@@ -22,10 +25,14 @@ def test_weight_loading(vllm_runner):
"""
Test parameter weight loading with tp>1.
"""
# MoE models need fp16.
NEEDS_FP16 = (QUANTIZATION == "gptq" or MODEL_NAME
== "nm-testing/test-w4a16-mixtral-actorder-group")
with vllm_runner(
model_name=MODEL_NAME,
revision=REVISION,
dtype=torch.half if QUANTIZATION == "gptq" else "auto",
dtype=torch.half if NEEDS_FP16 else "auto",
quantization=None if QUANTIZATION == "None" else QUANTIZATION,
max_model_len=MAX_MODEL_LEN,
tensor_parallel_size=2) as model: