[Cherry-Pick][CI] Use GPU-Build-RL runner for _build_linux_rl.yml(#7186) (#7197)

Author: EmmonsCurse

.github/workflows/_build_linux_rl.yml

@@ -8,7 +8,7 @@ on:
         description: "Build Images"
         required: true
         type: string
-        default: "iregistry.baidu-int.com/tiangexiao/base-images:paddlecloud-ubuntu24.04-gcc13.3-cuda12.9-cudnn9.9-bccl1.4.1.4-nccl2.26.5-openmpi4.1.5-FleetY13.0.0-rc2"
+        default: "iregistry.baidu-int.com/new_rl_infra/base-images:paddlecloud-ubuntu24.04-gcc13.3-cuda12.9-cudnn9.9-bccl1.4.1.4-nccl2.26.5-openmpi4.1.5-FleetY13.0.0-v2.4.0-rc1"
       FASTDEPLOY_ARCHIVE_URL:
         description: "URL of the compressed FastDeploy code archive."
         required: true
@@ -54,7 +54,7 @@ on:
         value: ${{ jobs.fd-build-rl.outputs.wheel_path_rl }}
 jobs:
   fd-build-rl:
-    runs-on: [self-hosted, GPU-Build]
+    runs-on: [self-hosted, GPU-Build-RL]
     timeout-minutes: 360
     outputs:
       wheel_path_rl: ${{ steps.set_output.outputs.wheel_path_rl }}
@@ -161,6 +161,7 @@ jobs:
             chown -R $(whoami) /workspace/FastDeploy
             cd FastDeploy
 
+            python -m pip uninstall paddlepaddle-gpu -y || true
             wget -q --no-proxy https://paddle-qa.bj.bcebos.com/paddle-pipeline/Release-TagBuild-Training-Linux-Gpu-Cuda12.9-Cudnn9.9-Trt10.5-Mkl-Avx-Gcc11-SelfBuiltPypiUse/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
             python -m pip install paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
 

.github/workflows/ce_job.yml

@@ -167,7 +167,7 @@ jobs:
       COMPILE_ARCH: "80,90"
       WITH_NIGHTLY_BUILD: OFF
       FD_VERSION: 0.0.0
-      PADDLE_WHL_URL: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Paddle-RL-Compile/develop/latest/paddlepaddle_gpu-3.3.0.dev-cp310-cp310-linux_x86_64.whl
+      PADDLE_WHL_URL: https://paddle-qa.bj.bcebos.com/paddle-pipeline/Release-TagBuild-Training-Linux-Gpu-Cuda12.9-Cudnn9.9-Trt10.5-Mkl-Avx-Gcc11-SelfBuiltPypiUse/latest/paddlepaddle_gpu-0.0.0-cp310-cp310-linux_x86_64.whl
 
   build_sm8689:
     name: BUILD_SM8689