Mixed-precision training has emerged as an indispensable tool for enhancing the efficiency of neural network training in recent years. Concurrently, JAX has grown in popularity as a versatile machine learning toolbox. However, it currently lacks robust support for mixed-precision training. We propose MPX, a mixed-precision training toolbox for JAX that simplifies and accelerates the training of large-scale neural networks while preserving model accuracy. MPX seamlessly integrates with popular toolboxes such as Equinox and Flax, allowing users to convert full-precision pipelines to mixed-precision versions with minimal modifications. By casting both inputs and outputs to half precision, and introducing a dynamic loss-scaling mechanism, MPX alleviates issues like gradient underflow and overflow that commonly arise in half precision computations. Its design inherits critical features from JAX's type-promotion behavior, ensuring that operations take place in the correct precision and allowing for selective enforcement of full precision where needed (e.g., sums, means, or softmax). MPX further provides wrappers for automatic creation and management of mixed-precision gradients and optimizers, enabling straightforward integration into existing JAX training pipelines. MPX's source code, documentation, and usage examples are available at github.com/Data-Science-in-Mechanical-Engineering/mixed_precision_for_JAX .

Exploiting low-precision computations has become a standard strategy in deep learning to address the growing computational costs imposed by ever larger models and datasets. However, naively performing all computations in low precision can lead to roundoff errors and instabilities. Therefore, mixed precision training schemes usually store the weights in high precision and use low-precision computations only for whitelisted operations. Despite their success, these principles are currently not reliable for training continuous-time architectures such as neural ordinary differential equations (Neural ODEs). This paper presents a mixed precision training framework for neural ODEs, combining explicit ODE solvers with a custom backpropagation scheme, and demonstrates its effectiveness across a range of learning tasks. Our scheme uses low-precision computations for evaluating the velocity, parameterized by the neural network, and for storing intermediate states, while stability is provided by a custom dynamic adjoint scaling and by accumulating the solution and gradients in higher precision. These contributions address two key challenges in training neural ODE: the computational cost of repeated network evaluations and the growth of memory requirements with the number of time steps or layers. Along with the paper, we publish our extendable, open-source PyTorch package rampde, whose syntax resembles that of leading packages to provide a drop-in replacement in existing codes. We demonstrate the reliability and effectiveness of our scheme using challenging test cases and on neural ODE applications in image classification and generative models, achieving approximately 50% memory reduction and up to 2x speedup while maintaining accuracy comparable to single-precision training.

As the parameters of Large Language Models (LLMs) have scaled to hundreds of billions, the demand for efficient training methods -- balancing faster computation and reduced memory usage without sacrificing accuracy -- has become more critical than ever. In recent years, various mixed precision strategies, which involve different precision levels for optimization components, have been proposed to increase training speed with minimal accuracy degradation. However, these strategies often require manual adjustments and lack theoretical justification. In this work, we leverage stochastic rounding (SR) to address numerical errors of training with low-precision representation. We provide theoretical analyses of implicit regularization and convergence under the Adam optimizer when SR is utilized. With the insights from these analyses, we extend previous BF16 + SR strategy to be used in distributed settings, enhancing the stability and performance for large scale training. Empirical results from pre-training models with up to 6.7B parameters, for the first time, demonstrate that our BF16 with SR strategy outperforms (BF16, FP32) mixed precision strategies, achieving better validation perplexity, up to $1.54\times$ higher throughput, and $30\%$ less memory usage.

Mixed precision training combines half-precision (FP16) and single-precision (FP32) computations This project introduces SplaXBERT, a questionanswering to improve efficiency while preserving accuracy.

Floating-point formats are not the most glamorous or (frankly) the important consideration when working with deep learning models: if your model isn't working well, then your floating-point format certainly isn't going to save you! However, past a certain point of model complexity/model size/training time, your choice of floating-point format can have a significant impact on your model training times and even performance. Here's how the rest of this post is structured: It's more likely than you think! It's been known for quite some time that deep neural networks can tolerate lower numerical precision. High-precision calculations turn out not to be that useful in training or inferencing neural networks: the additional precision confers no benefit while being slower and less memory-efficient.

With the increasing size of deep learning models, the memory and compute demands too have increased. Techniques have been developed to train deep neural networks faster. One approach is to use half-precision floating-point numbers; FP16 instead of FP32. Recently, researchers have found that using them together is a smarter choice. Mixed precision is one such technique, which can be used to train with half-precision while maintaining the network accuracy achieved with single precision.

Image classification is the Hello World of deep learning. For me, that project was Pneumonia Detection using Chest X-rays. Since this was a relatively small dataset, I could train my model in about 50 minutes. The dataset I worked with, involved around 4,500 images. And the only reason it took 50 minutes was because the images were high definition.

This repository provides scripts to train the Jasper model to achieve near state of the art accuracy and perform high-performance inference using NVIDIA TensorRT. This repository is tested and maintained by NVIDIA. The Jasper model is an end-to-end neural acoustic model for automatic speech recognition (ASR) that provides near state-of-the-art results on LibriSpeech among end-to-end ASR models without any external data. The Jasper architecture of convolutional layers was designed to facilitate fast GPU inference, by allowing whole sub-blocks to be fused into a single GPU kernel. This is important for meeting strict real-time requirements of ASR systems in deployment. The results of the acoustic model are combined with the results of external language models to get the top-ranked word sequences corresponding to a given audio segment.

Recent work has shown that quantization can help reduce the memory, compute, and energy demands of deep neural networks without significantly harming their quality. However, whether these prior techniques, applied traditionally to image-based models, work with the same efficacy to the sequential decision making process in reinforcement learning remains an unanswered question. To address this void, we conduct the first comprehensive empirical study that quantifies the effects of quantization on various deep reinforcement learning policies with the intent to reduce their computational resource demands. We apply techniques such as post-training quantization and quantization aware training to a spectrum of reinforcement learning tasks (such as Pong, Breakout, BeamRider and more) and training algorithms (such as PPO, A2C, DDPG, and DQN). Across this spectrum of tasks and learning algorithms, we show that policies can be quantized to 6-8 bits of precision without loss of accuracy. We also show that certain tasks and reinforcement learning algorithms yield policies that are more difficult to quantize due to their effect of widening the models' distribution of weights and that quantization aware training consistently improves results over post-training quantization and oftentimes even over the full precision baseline. Finally, we demonstrate real-world applications of quantization for reinforcement learning. We use half-precision training to train a Pong model 50% faster, and we deploy a quantized reinforcement learning based navigation policy to an embedded system, achieving an 18$\times$ speedup and a 4$\times$ reduction in memory usage over an unquantized policy.

Today the Apache MXNet community is announcing the 1.2.0 release of the Apache MXNet deep learning framework -- faster, easier to use: ONNX, MKL-DNN support, mixed precision training and many more features! Check the full release note here. MKL-DNN integration: MXNet now integrates with Intel MKL-DNN to accelerate neural network operators: Convolution, Deconvolution, FullyConnected, Pooling, Batch Normalization, Activation, LRN, Softmax, as well as some common operators like sum and concat. This integration allows NDArray to contain data with MKL-DNN layouts and reduces data layout conversion overhead to get the maximal performance from MKL-DNN. More details are available in this blog post.