OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

deeprec.readthedocs.io

deeprec.readthedocs.io hosts Chinese-language documentation for DeepRec, a deep-learning framework focused on large-scale sparse models. The docs describe distributed training, graph and runtime optimization, embedding features, source compilation and a processor library for prediction integrations.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

deeprec.readthedocs.io — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use the documentation when assessing DeepRec for sparse-model infrastructure or when preparing a source build with the listed compiler, Python and CUDA combinations.

Who it is for

  • Machine-learning engineers building large-scale sparse models, including recommendation systems.
  • Infrastructure developers planning CPU or GPU source builds and distributed training workflows.
  • Teams evaluating embedding optimization and processor-library integration for prediction.

Documented uses

  • Compile DeepRec from source into a pip wheel for a selected CPU or GPU environment.
  • Plan distributed training for sparse-model workloads.
  • Evaluate embedding approaches for high-cardinality features.
  • Build the processor shared library and use its serving API for prediction.
Before you choose

Requirements & limits

The documentation site is active and accessible, and the technical pages are in Chinese.

  • The available build instructions are specific to listed GCC, Python, Docker and CUDA combinations.
  • GPU compilation requires choosing compute-capability values suited to the target hardware.
  • The DeepRec-specific Estimator instructions address distributed protocols including grpc++ and star_server.
  • Processor-library compilation and framework compilation are documented as separate build paths.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Compile DeepRec into a pip package

Build a CPU or GPU DeepRec pip package from source using a documented environment and Bazel target.

Before you start

  • A listed DeepRec CPU or GPU development Docker image, matching its GCC and Python versions.
  • For GPU builds, the relevant CUDA environment and a configured TF_CUDA_COMPUTE_CAPABILITIES value for the target GPU architecture.
  • Bazel and the DeepRec source tree configured for the chosen build.

Steps

  1. Choose a documented CPU or GPU development image and match the compiler, Python and, for GPU builds, CUDA versions.

  2. For GPU builds, set TF_CUDA_COMPUTE_CAPABILITIES to the required architecture values and run ./configure.

  3. Run the documented Bazel pip-package build target: bazel build -c opt --config=opt //tensorflow/tools/pip_package:build_pip_package.

  4. Generate a wheel with ./bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/tensorflow_pkg.

Expected outcome

A DeepRec wheel package is generated in the selected output directory.

Guide limitations

  • The guide lists distinct environment combinations; the selected image and build options must match the intended CPU or GPU target.
  • The documented wheel-install command uses a specific Python 3.8 Linux wheel filename, so it does not establish installation compatibility for other environments.
Documentation 1 ↗
Evidence and transparency

Sources & review

deeprec.readthedocs.io ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
12 October 2026
Original website title
简介 — DeepRec latest 文档 · Official homepage title · retained in its source language
How we review sourcesSuggest a correction