Skip to content

PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度

Increase memory limits for the build process in the CI pipeline (#2198)

chen, suyue8ae2b2e13cdd
diff --git a/.azure-pipelines/template/lib-build-template.yml b/.azure-pipelines/template/lib-build-template.yml
index 711ff6754..ced6c1abf 100644
--- a/.azure-pipelines/template/lib-build-template.yml
+++ b/.azure-pipelines/template/lib-build-template.yml
@@ -91,6 +91,8 @@ stages:
-e HOME=/tmp \
--cpuset-cpus="${NUMA_CPUSET:-0-15}" \
--cpuset-mems="${NUMA_NODE:-0}" \
+ --memory="128g" \
+ --memory-swap="128g" \
-v $(Build.SourcesDirectory):/auto-round \
-v /opt/intel/oneapi:/opt/intel/oneapi \
-w /auto-round/auto_round_extension/ark \