Use mirror images (#1912)

### What this PR does / why we need it?
More discussion can be found
[here](https://github.com/ascend-gha-runners/docs/issues/23).

The infra team deployed a internal registry since both `m.daocloud.io`
and `quay.io` suffered a unstable connect quality.

CI will benefit both the connection and download speed by switching to
the internal registry.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
tested locally

- vLLM version: v0.9.2
- vLLM main:
6b46c4b653

---------

Signed-off-by: mywaaagh_admin <pkwarcraft@gmail.com>

.github/workflows/accuracy_test.yaml

@@ -123,7 +123,7 @@ jobs:
       fail-fast: false
     name: ${{ matrix.model_name }} accuracy
     container:
-      image: m.daocloud.io/quay.io/ascend/cann:8.1.rc1-910b-ubuntu22.04-py3.10
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.1.rc1-910b-ubuntu22.04-py3.10
       env:
         DATASET_SOURCE: ModelScope
         VLLM_USE_MODELSCOPE: True