Installing Dependencies
To ensure the normal use of RAG SDK, you need to install related dependencies.
- If RAG SDK is deployed on a physical machine, all dependencies in Table 1 need to be installed.
- If RAG SDK is deployed in a container, install the npu-driver package, npu-firmware package, and Ascend Docker Runtime on the host and start the MindIE inference service. For details about other procedures, see Deploying RAG SDK in a Container.
Installing Software Packages
- Install the NPU driver and firmware. For details, see "Installing the NPU Driver and Firmware" in CANN Software Installation Guide (Commercial Edition) or "Installing the NPU Driver and Firmware" in CANN Software Installation Guide (Community Edition). To enable non-root users to use the driver, add the --install-for-all option to the driver installation.
- Optional: Install CANN ToolKit, ops, and NNAL acceleration library by referring to "Installing Dependencies" and "Installing the CANN Package" in CANN Software Installation Guide. You are advised to install them as the common user HwHiAiUser. If RAG SDK is deployed using the AscendHub image, skip this step.
- Install and run a large inference model. For details, see "Mode 3: Container Installation" and "Configuring Server" in MindIE Installation Guide.
- Install Ascend Docker Runtime. For details, see "Installation" > "Installation and Deployment" in MindCluster Cluster Scheduling User Guide.
Downloading Software Packages
Short Name |
Full Name |
Version |
Link |
|---|---|---|---|
CANN package |
Ascend-cann-toolkit_<version>_linux-<arch>.run |
8.5.0 |
|
ops (CANN operator package) |
Ascend-cann-<chip_type>-ops_<version>_linux-<arch>.run |
||
(Optional) NNAL acceleration library |
Ascend-cann-nnal_<version>_linux-<arch>.run |
||
npu-driver |
Ascend-hdk-<chip_type>-npu-driver_<version>_linux-<arch>.run |
25.5.0 |
|
npu-firmware |
Ascend-hdk-<chip_type>-npu-firmware_<version>.run |
||
Index SDK package |
Ascend-mindxsdk-mxindex_<version>_linux-<arch>.run |
7.3.0 |
|
MindIE inference engine package |
Ascend-mindie_<version>_linux-<arch>.run |
2.3.0 |
|
Ascend Docker Runtime |
Ascend-docker-runtime_<version>_linux-<arch>.run |
7.3.0 |
|
Python |
- |
3.11/3.12 |
- <version> indicates the software version.
- <arch> indicates the CPU architecture.
- <chip_type> indicates the processor type. You can run the npu-smi info command on the server where the Ascend AI Processor is installed and then delete the last digit of Name. The obtained value is the value of <chip_type>.
- To enable a non-root user to use the driver, add the --install-for-all option when installing the npu-driver.
- For open-source and third-party software integrated by users, check the vulnerabilities and issues in the corresponding community and fix them in a timely manner. You can confirm the known vulnerabilities of the corresponding open-source software version on the CVE official website, and fix the vulnerabilities through version upgrade or patch package update.
Parent topic: Installation and Deployment