Category: Blog

Pruning Hugging Face BERT: Apply both pruning and layer dropping sparsification methods to increase BERT performance anywhere from 3.3x to 14x on CPUs depending on accuracy constraints In this post, we go into detail on pruning Hugging Face BERT and describe how sparsification combined with the DeepSparse Engine improves BERT model performance on CPUs. We’ll… Read More Pruning Hugging Face BERT: Using Compound Sparsification for Faster CPU Inference with Better Accuracy

Neural Magic has been busy this summer on the Community Edition (CE) of our DeepSparse tools; we’re excited to share highlights of releases 0.5 and 0.6. The full technical release notes are always found within our GitHub release indexes linked from our Docs website or the specific Neural Magic repository. For user help or questions… Read More Neural Magic CE 0.5 and 0.6 Product Releases

Prune and quantize YOLOv5 for a 10x increase in performance with 12x smaller model files. Neural Magic improves YOLOv5 model performance on CPUs by using state-of-the-art pruning and quantization techniques combined with the DeepSparse Engine. In this blog post, we’ll cover our general methodology and demonstrate how to: Leverage the Ultralytics YOLOv5 repository with SparseML’s sparsification… Read More YOLOv5 on CPUs: Sparsifying to Achieve GPU-Level Performance and a Smaller Footprint

Sparsifying YOLOv3 (or any other model) involves removing redundant information from neural networks using algorithms such as pruning and quantization, among others. This sparsification process results in many benefits for deployment environments, including faster inference and smaller file sizes. Unfortunately, many have not realized the benefits due to the complicated process and number of hyperparameters… Read More New Tutorial: Sparsifying YOLOv3 Using Recipes

We are excited to announce that industry veteran Brian Stevens will be joining Neural Magic as Chief Executive Officer. Brian brings vast experience in open source, enterprise, and hyper-scale cloud to the team. Before joining Neural Magic, Brian was Vice President and CTO of Google Cloud and Executive Vice President and CTO of Red Hat.… Read More Neural Magic Appoints Brian Stevens as Chief Executive Officer

Use CPUs to decrease costs and increase deployment flexibility while still achieving GPU-class performance. In this post, we elaborate on how we used state-of-the-art pruning and quantization techniques to improve the performance of the YOLOv3 on CPUs. We’ll show that by leveraging the robust YOLO training framework from Ultralytics with SparseML’s sparsification recipes it is… Read More YOLOv3 on CPUs: Sparsifying to Achieve GPU-Level Performance

To understand how Neural Magic’s Deep Sparse technology works, it’s important to quickly cover the journey of our founders. While mapping the neural connections in the brain at MIT, Neural Magic’s founders Nir Shavit and Alexander Matveev were frustrated with the many limitations imposed by GPUs. Along the way, they stopped to ask themselves a… Read More How Neural Magic’s Deep Sparse Technology Works

Did you know that most weights in a neural network are actually useless? In other words, most weights can be removed with little to no impact on the loss. But, how and why would you optimize a deep learning model in practice? Through a combination of pruning and quantization (or “sparse-quantization”) you can drastically improve… Read More Using Sparse-Quantization in Inference: NeurIPS 2020