Apache MXNet

A deep learning software framework, used to train, and deploy deep neural networks

⚖️ Free · Open

Features & Limitations

+	Multi-Language Support	Supports multiple programming languages including Python, R, Julia, and Go.
+	Hybrid Programming	Combines declarative symbolic expression with imperative tensor computation
+	Auto Differentiation	Offers automatic differentiation to derive gradients
+	Efficient Execution	Runs efficiently on various systems, from mobile devices to GPU clusters
+	Optimization	Provides global computation graph for performance and memory optimization
+	KVStore	Facilitates data synchronization over devices with a distributed key-value store
+	Graph Optimization	Employs strategies like memory allocation and operation grouping for efficiency
+	Scalable	Designed for scalability across distributed systems for large-scale deep neural network applications
-	Community Size	Smaller open-source community compared to TensorFlow.
-	Delayed Improvements	Longer time for improvements and bug fixes due to limited community support.
-	Data-Parallelism	Supports data-parallelism only (single GPU model size limitation)
-	Sparse Model Incompatibility	Not friendly to sparse models; BytePS focuses on dense DNN models.
-	Scala Conversion Overhead	Overhead in converting collections when using Scala
-	Increased JAR File Size	Larger JAR files due to Scala inclusion

Not available, but we appreciate help! You can help us improve this page by contacting us.

Not available, but we appreciate help! You can help us improve this page by contacting us.

C++, Python, Perl, Scala, CUDA, R, JavaScript, Julia, Golang

8 April 2014