Installation Guide

System Requirements

Python Version - Python 3.10+

Dependencies - PyTorch 2.4+ - CUDA 12.4 or ROCm 7.0+

Installation Methods

Docker Installation (Recommended)

In NVIDIA GPU, we provide build scripts for three versions of images: pytorch240/251/260. You can build the base image using commands, for example:

Clone the repository:

git clone https://github.com/alibaba/RecIS.git
cd recis
git submodule update --init --recursive

Build Docker image for pytorch240 version:

docker build --network=host -f docker/Dockerfile.torch240 -t recis:torch240 .

Start the torch240 version image:

docker run --runtime=nvidia --net=host -it --cpuset-cpus="0-63" -m 300G recis:torch240 /bin/bash

Source Installation

1. Clone the repository:

   git clone https://github.com/alibaba/RecIS.git
   cd recis
   git submodule update --init --recursive
   

2. Build and install RecIS:

   bash build.sh 0
   pip install `find ./dist -name "recis*.whl" -maxdepth 1`
   

3. Build and install column-io:

   cd third_party/column-io/
   bash tools/build_and_install.sh 0
   

4. Verify installation:

   python -c "import recis; print('RecIS installed successfully!')"
   

Installing on AMD GPUs

RecIS also supports AMD GPUs, with ROCm 7.0.2 verified. Similar to NVIDIA GPUs, you can install RecIS via Docker or from source.

Build the Docker image for rocm7.0.2-pytorch271:

docker build --network=host -f docker/Dockerfile.rocm.torch271 -t recis:rocm702_torch271 .

Start the rocm7.0.2-pytorch271 Docker image:

docker run -it \
    --cap-add=SYS_PTRACE \
    --security-opt seccomp=unconfined \
    --device=/dev/kfd \
    --device=/dev/dri \
    --group-add video \
    --ipc=host \
    --shm-size 8G \
    recis:rocm702_torch271 /bin/bash

When installing RecIS from source on AMD GPUs, note that libhipcxx is not included with the ROCm SDK and must be manually built and installed in your environment first:

cd ~ && git clone https://github.com/ROCm/libhipcxx.git -b release/2.2.x
cd libhipcxx/ && mkdir build && cd build && cmake .. -DCMAKE_INSTALL_PREFIX=/opt/rocm && make install
rm -rf libhipcxx/

Installation Verification

Run the following code to verify successful installation:

import torch
import recis

# Check versions
print(f"PyTorch version: {torch.__version__}")

# Check GPU support
print(f"CUDA available: {torch.cuda.is_available()}")
if torch.cuda.is_available():
    print(f"CUDA version: {torch.version.cuda}") # or print(f"HIP version: {torch.version.hip}")
    print(f"GPU count: {torch.cuda.device_count()}")

# Simple functionality test
from recis.nn.modules.embedding import DynamicEmbedding, EmbeddingOption

emb_opt = EmbeddingOption(embedding_dim=16)
emb = DynamicEmbedding(emb_opt)
print("RecIS core modules loaded successfully!")

If the above code runs without errors, RecIS has been successfully installed.