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

Improve configs - the rest! (#17562)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-05-09 23:18:44 +01:00
committed by GitHub
parent 7e3571134f
commit 4b2ed7926a
14 changed files with 456 additions and 340 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/compile/test_silu_mul_quant_fusion.py
View File

@@ -6,7 +6,7 @@ import vllm.envs as envs
from vllm._custom_ops import scaled_fp8_quant
from vllm.compilation.activation_quant_fusion import ActivationQuantFusionPass
from vllm.compilation.fx_utils import find_auto_fn, find_auto_fn_maybe
from vllm.config import CompilationConfig, VllmConfig
from vllm.config import CompilationConfig, PassConfig, VllmConfig
from vllm.model_executor.layers.activation import SiluAndMul
from .backend import TestBackend
@@ -36,8 +36,7 @@ def test_fusion_silu_and_mul_quant(num_tokens, hidden_size):
# Reshape pass is needed for the fusion pass to work
config = VllmConfig()
config.compilation_config = CompilationConfig(
pass_config=CompilationConfig.PassConfig(enable_fusion=True,
enable_reshape=True))
pass_config=PassConfig(enable_fusion=True, enable_reshape=True))
fusion_pass = ActivationQuantFusionPass(config)
backend = TestBackend(fusion_pass)