#7272nd of 75 in ML frameworks and libraries
deeplearning4j
JVM deep learning stack with ND4J tensors, SameDiff autodiff and DataVec ETL
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 14.3k
- License
- Apache-2.0
- Last commit
- Feb 2026
- Language
- Java
14,137 on 31 Mar 202614,274 on 20 Oct 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Deeplearning4j is a set of JVM libraries for building and training neural networks in Java, Scala or Kotlin. It includes a high-level network API (DL4J), the ND4J array library backed by C++ LibND4J with CPU and CUDA support, the SameDiff autodiff framework, and DataVec for data loading. It imports Keras h5 models and TensorFlow frozen .pb graphs, and trains on Apache Spark.
Who it is for: JVM developers building or training neural networks in Java or Scala
Strengths
- Runs natively on the JVM; usable from Java, Scala, Kotlin and Clojure
- ND4J backends cover x86 AVX2/512, ARM, ppc64le and CUDA GPUs
- Imports Keras h5 and TensorFlow frozen .pb models
- Distributed training on Apache Spark; DataVec reads HDFS, CSV, image and audio data
Weaknesses
- No release listed; README pins 1.0.0-M2.1, a milestone version
- SameDiff import for ONNX, SavedModel and Keras is listed as planned
- SameDiff is graph-mode only; eager execution is planned
- Tests only run on JDK 11; README lists CUDA 10.x versions
What it needs
- GPU optional
- Needs Maven or other JVM build tool, JavaCPP, Apache Spark (optional)
- Models: Keras h5, TensorFlow frozen .pb
Also in ML frameworks and libraries
See all 75| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd for deep learning | 94 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 92 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and reference training scripts | 91 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 90 out of 100 |