Search

Benchmarks are essential in the design of modern HPC installations, as they define key aspects of system components. Beyond synthetic workloads, it is crucial to include real applications that represent user requirements into benchmark suites, to guarantee high usability and widespread adoption of a new system. Given the significant investments in leadership-class supercomputers of the exascale era, this is even more important and necessitates alignment with a vision of Open Science and reproducibility. In this work, we present the JUPITER Benchmark Suite, which incorporates 16 applications from various domains. It was designed for and used in the procurement of JUPITER, the first European exascale supercomputer. We identify requirements and challenges and outline the project and software infrastructure setup. We provide descriptions and scalability studies of selected applications and a set of key takeaways. The JUPITER Benchmark Suite is released as open source software with this work at https://github.com/FZJ-JSC/jubench., Comment: To be published in Proceedings of The International Conference for High Performance Computing Networking, Storage, and Analysis (SC '24) (2024)

Concentrating Solar Power (CSP) plants play a crucial role in the global transition towards sustainable energy. A key factor in ensuring the safe and efficient operation of CSP plants is the distribution of concentrated flux density on the receiver. However, the non-ideal flux density generated by individual heliostats can undermine the safety and efficiency of the power plant. The flux density from each heliostat is influenced by its precise surface profile, which includes factors such as canting and mirror errors. Accurately measuring these surface profiles for a large number of heliostats in operation is a formidable challenge. Consequently, control systems often rely on the assumption of ideal surface conditions, which compromises both safety and operational efficiency. In this study, we introduce inverse Deep Learning Ray Tracing (iDLR), an innovative method designed to predict heliostat surfaces based solely on target images obtained during heliostat calibration. Our simulation-based investigation demonstrates that sufficient information regarding the heliostat surface is retained in the flux density distribution of a single heliostat, enabling deep learning models to accurately predict the underlying surface with deflectometry-like precision for the majority of heliostats. Additionally, we assess the limitations of this method, particularly in relation to surface accuracy and resultant flux density predictions. Furthermore, we are presenting a new comprehensive heliostat model using Non-Uniform Rational B-Spline (NURBS) that has the potential to become the new State of the Art for heliostat surface parameterization. Our findings reveal that iDLR has significant potential to enhance CSP plant operations, potentially increasing the overall efficiency and energy output of the power plants.

Kaplan et al. and Hoffmann et al. developed influential scaling laws for the optimal model size as a function of the compute budget, but these laws yield substantially different predictions. We explain the discrepancy by reproducing the Kaplan scaling law on two datasets (OpenWebText2 and RefinedWeb) and identifying three factors causing the difference: last layer computational cost, warmup duration, and scale-dependent optimizer tuning. With these factors corrected, we obtain excellent agreement with the Hoffmann et al. (i.e., "Chinchilla") scaling law. Counter to a hypothesis of Hoffmann et al., we find that careful learning rate decay is not essential for the validity of their scaling law. As a secondary result, we derive scaling laws for the optimal learning rate and batch size, finding that tuning the AdamW $\beta_2$ parameter is essential at lower batch sizes., Comment: NeurIPS 2024 revisions

We introduce DataComp for Language Models (DCLM), a testbed for controlled dataset experiments with the goal of improving language models. As part of DCLM, we provide a standardized corpus of 240T tokens extracted from Common Crawl, effective pretraining recipes based on the OpenLM framework, and a broad suite of 53 downstream evaluations. Participants in the DCLM benchmark can experiment with data curation strategies such as deduplication, filtering, and data mixing at model scales ranging from 412M to 7B parameters. As a baseline for DCLM, we conduct extensive experiments and find that model-based filtering is key to assembling a high-quality training set. The resulting dataset, DCLM-Baseline enables training a 7B parameter language model from scratch to 64% 5-shot accuracy on MMLU with 2.6T training tokens. Compared to MAP-Neo, the previous state-of-the-art in open-data language models, DCLM-Baseline represents a 6.6 percentage point improvement on MMLU while being trained with 40% less compute. Our baseline model is also comparable to Mistral-7B-v0.3 and Llama 3 8B on MMLU (63% & 66%), and performs similarly on an average of 53 natural language understanding tasks while being trained with 6.6x less compute than Llama 3 8B. Our results highlight the importance of dataset design for training language models and offer a starting point for further research on data curation., Comment: Project page: https://www.datacomp.ai/dclm/

Large Language Models (LLMs) are often described as being instances of foundation models - that is, models that transfer strongly across various tasks and conditions in few-show or zero-shot manner, while exhibiting scaling laws that predict function improvement when increasing the pre-training scale. These claims of excelling in different functions and tasks rely on measurements taken across various sets of standardized benchmarks showing high scores for such models. We demonstrate here a dramatic breakdown of function and reasoning capabilities of state-of-the-art models trained at the largest available scales which claim strong function, using a simple, short, conventional common sense problem (AIW problem) formulated in concise natural language, easily solvable by humans. The breakdown is dramatic, as models show strong fluctuations across even slight problem variations that should not affect problem solving, also expressing strong overconfidence in the wrong solutions, often backed up by plausible sounding explanation-like confabulations. Various standard interventions in an attempt to get the right solution, like various type of enhanced prompting, or urging the models to reconsider the wrong solutions again by multi step re-evaluation, fail. We take these initial observations to the scientific and technological community to stimulate urgent re-assessment of the claimed capabilities of current generation of LLMs. Such re-assessment also requires common action to create standardized benchmarks that would allow proper detection of such basic reasoning deficits that obviously manage to remain undiscovered by current state-of-the-art evaluation procedures and benchmarks. Code for reproducing experiments in the paper and raw experiments data can be found at https://github.com/LAION-AI/AIW, Comment: v2.01. Minor edits. Further experiments on various AIW problem variations. AIW "Alice Female Power Boost", AIW Extension (AIW Ext). Including recent Claude 3.5 Sonnet and Qwen 2 72B Instruct results

Scaling laws are useful guides for derisking expensive training runs, as they predict performance of large models using cheaper, small-scale experiments. However, there remain gaps between current scaling studies and how language models are ultimately trained and evaluated. For instance, scaling is usually studied in the compute-optimal training regime (i.e., "Chinchilla optimal" regime). In contrast, models are often over-trained to reduce inference costs. Moreover, scaling laws mostly predict loss on next-token prediction, but models are usually compared on downstream task performance. To address both shortcomings, we create a testbed of 104 models with 0.011B to 6.9B parameters trained with various numbers of tokens on three data distributions. First, we fit scaling laws that extrapolate in both the amount of over-training and the number of model parameters. This enables us to predict the validation loss of a 1.4B parameter, 900B token run (i.e., 32$\times$ over-trained) and a 6.9B parameter, 138B token run (i.e., a compute-optimal run)$\unicode{x2014}$each from experiments that take 300$\times$ less compute. Second, we relate the perplexity of a language model to its downstream task performance by proposing a power law. We use this law to predict top-1 error averaged over downstream tasks for the two aforementioned models, using experiments that take 20$\times$ less compute. Our experiments are available at https://github.com/mlfoundations/scaling.

Over 1.4 billion chest X-rays (CXRs) are performed annually due to their cost-effectiveness as an initial diagnostic test. This scale of radiological studies provides a significant opportunity to streamline CXR interpretation and documentation. While foundation models are a promising solution, the lack of publicly available large-scale datasets and benchmarks inhibits their iterative development and real-world evaluation. To overcome these challenges, we constructed a large-scale dataset (CheXinstruct), which we utilized to train a vision-language foundation model (CheXagent). We systematically demonstrated competitive performance across eight distinct task types on our novel evaluation benchmark (CheXbench). Beyond technical validation, we assessed the real-world utility of CheXagent in directly drafting radiology reports. Our clinical assessment with eight radiologists revealed a 36% time saving for residents using CheXagent-drafted reports, while attending radiologists showed no significant time difference editing resident-drafted or CheXagent-drafted reports. The CheXagent-drafted reports improved the writing efficiency of both radiology residents and attending radiologists in 81% and 61% of cases, respectively, without loss of quality. Overall, we demonstrate that CheXagent can effectively perform a variety of CXR interpretation tasks and holds potential to assist radiologists in routine clinical workflows., Comment: 26 pages, 8 figures

We introduce OpenFlamingo, a family of autoregressive vision-language models ranging from 3B to 9B parameters. OpenFlamingo is an ongoing effort to produce an open-source replication of DeepMind's Flamingo models. On seven vision-language datasets, OpenFlamingo models average between 80 - 89% of corresponding Flamingo performance. This technical report describes our models, training data, hyperparameters, and evaluation suite. We share our models and code at https://github.com/mlfoundations/open_flamingo.

Multimodal datasets are a critical component in recent breakthroughs such as Stable Diffusion and GPT-4, yet their design does not receive the same research attention as model architectures or training algorithms. To address this shortcoming in the ML ecosystem, we introduce DataComp, a testbed for dataset experiments centered around a new candidate pool of 12.8 billion image-text pairs from Common Crawl. Participants in our benchmark design new filtering techniques or curate new data sources and then evaluate their new dataset by running our standardized CLIP training code and testing the resulting model on 38 downstream test sets. Our benchmark consists of multiple compute scales spanning four orders of magnitude, which enables the study of scaling trends and makes the benchmark accessible to researchers with varying resources. Our baseline experiments show that the DataComp workflow leads to better training sets. In particular, our best baseline, DataComp-1B, enables training a CLIP ViT-L/14 from scratch to 79.2% zero-shot accuracy on ImageNet, outperforming OpenAI's CLIP ViT-L/14 by 3.7 percentage points while using the same training procedure and compute. We release DataComp and all accompanying code at www.datacomp.ai., Comment: NeurIPS 2023 Datasets and Benchmarks Track

Solar activity is one of the main drivers of variability in our solar system and the key source of space weather phenomena that affect Earth and near Earth space. The extensive record of high resolution extreme ultraviolet (EUV) observations from the Solar Dynamics Observatory (SDO) offers an unprecedented, very large dataset of solar images. In this work, we make use of this comprehensive dataset to investigate capabilities of current state-of-the-art generative models to accurately capture the data distribution behind the observed solar activity states. Starting from StyleGAN-based methods, we uncover severe deficits of this model family in handling fine-scale details of solar images when training on high resolution samples, contrary to training on natural face images. When switching to the diffusion based generative model family, we observe strong improvements of fine-scale detail generation. For the GAN family, we are able to achieve similar improvements in fine-scale generation when turning to ProjectedGANs, which uses multi-scale discriminators with a pre-trained frozen feature extractor. We conduct ablation studies to clarify mechanisms responsible for proper fine-scale handling. Using distributed training on supercomputers, we are able to train generative models for up to 1024x1024 resolution that produce high quality samples indistinguishable to human experts, as suggested by the evaluation we conduct. We make all code, models and workflows used in this study publicly available at \url{https://github.com/SLAMPAI/generative-models-for-highres-solar-images}.

Scaling up neural networks has led to remarkable performance across a wide range of tasks. Moreover, performance often follows reliable scaling laws as a function of training set size, model size, and compute, which offers valuable guidance as large-scale experiments are becoming increasingly expensive. However, previous work on scaling laws has primarily used private data \& models or focused on uni-modal language or vision learning. To address these limitations, we investigate scaling laws for contrastive language-image pre-training (CLIP) with the public LAION dataset and the open-source OpenCLIP repository. Our large-scale experiments involve models trained on up to two billion image-text pairs and identify power law scaling for multiple downstream tasks including zero-shot classification, retrieval, linear probing, and end-to-end fine-tuning. We find that the training distribution plays a key role in scaling laws as the OpenAI and OpenCLIP models exhibit different scaling behavior despite identical model architectures and similar training recipes. We open-source our evaluation workflow and all models, including the largest public CLIP models, to ensure reproducibility and make scaling laws research more accessible. Source code and instructions to reproduce this study will be available at https://github.com/LAION-AI/scaling-laws-openclip, Comment: CVPR 2023. Version with minor extension. Original: https://openaccess.thecvf.com/content/CVPR2023/html/Cherti_Reproducible_Scaling_Laws_for_Contrastive_Language-Image_Learning_CVPR_2023_paper

In this paper, deep learning (DL) methods are evaluated in the context of turbulent flows. Various generative adversarial networks (GANs) are discussed with respect to their suitability for understanding and modeling turbulence. Wasserstein GANs (WGANs) are then chosen to generate small-scale turbulence. Highly resolved direct numerical simulation (DNS) turbulent data is used for training the WGANs and the effect of network parameters, such as learning rate and loss function, is studied. Qualitatively good agreement between DNS input data and generated turbulent structures is shown. A quantitative statistical assessment of the predicted turbulent fields is performed.

Groundbreaking language-vision architectures like CLIP and DALL-E proved the utility of training on large amounts of noisy image-text data, without relying on expensive accurate labels used in standard vision unimodal supervised learning. The resulting models showed capabilities of strong text-guided image generation and transfer to downstream tasks, while performing remarkably at zero-shot classification with noteworthy out-of-distribution robustness. Since then, large-scale language-vision models like ALIGN, BASIC, GLIDE, Flamingo and Imagen made further improvements. Studying the training and capabilities of such models requires datasets containing billions of image-text pairs. Until now, no datasets of this size have been made openly available for the broader research community. To address this problem and democratize research on large-scale multi-modal models, we present LAION-5B - a dataset consisting of 5.85 billion CLIP-filtered image-text pairs, of which 2.32B contain English language. We show successful replication and fine-tuning of foundational models like CLIP, GLIDE and Stable Diffusion using the dataset, and discuss further experiments enabled with an openly available dataset of this scale. Additionally we provide several nearest neighbor indices, an improved web-interface for dataset exploration and subset generation, and detection scores for watermark, NSFW, and toxic content detection. Announcement page https://laion.ai/laion-5b-a-new-era-of-open-large-scale-multi-modal-datasets/, Comment: 36th Conference on Neural Information Processing Systems (NeurIPS 2022), Track on Datasets and Benchmarks. OpenReview: https://openreview.net/forum?id=M3Y74vmsMcY

Multi-modal language-vision models trained on hundreds of millions of image-text pairs (e.g. CLIP, DALL-E) gained a recent surge, showing remarkable capability to perform zero- or few-shot learning and transfer even in absence of per-sample labels on target image data. Despite this trend, to date there has been no publicly available datasets of sufficient scale for training such models from scratch. To address this issue, in a community effort we build and release for public LAION-400M, a dataset with CLIP-filtered 400 million image-text pairs, their CLIP embeddings and kNN indices that allow efficient similarity search., Comment: Short version. Accepted at Data Centric AI NeurIPS Workshop 2021

In this article, we present JUWELS Booster, a recently commissioned high-performance computing system at the J\"ulich Supercomputing Center. With its system architecture, most importantly its large number of powerful Graphics Processing Units (GPUs) and its fast interconnect via InfiniBand, it is an ideal machine for large-scale Artificial Intelligence (AI) research and applications. We detail its system architecture, parallel, distributed model training, and benchmarks indicating its outstanding performance. We exemplify its potential for research application by presenting large-scale AI research highlights from various scientific fields that require such a facility., Comment: 12 pages, 5 figures. Accepted at ISC 2021, Workshop Deep Learning on Supercomputers. This is a duplicate submission as my previous submission is on hold for several weeks now and my attempts to contact the moderators failed

Increasing model, data and compute budget scale in the pre-training has been shown to strongly improve model generalization and transfer learning in vast line of work done in language modeling and natural image recognition. However, most studies on the positive effect of larger scale were done in scope of in-domain setting, with source and target data being in close proximity. To study effect of larger scale for both in-domain and out-of-domain setting when performing full and few-shot transfer, we combine here for the first time large, openly available medical X-Ray chest imaging datasets to reach a scale for medical imaging domain comparable to ImageNet-1k, routinely used for pre-training in natural image domain. We then conduct supervised pre-training, while varying network size and source data scale and domain, being either large natural (ImageNet-1k/21k) or large medical chest X-Ray datasets, and transfer pre-trained models to different natural or medical targets. We observe strong improvement due to larger pre-training scale for intra-domain natural-natural and medical-medical transfer. For inter-domain natural-medical transfer, we find improvements due to larger pre-training scale on larger X-Ray targets in full shot regime, while for smaller targets and for few-shot regime the improvement is not visible. Remarkably, large networks pre-trained on very large natural ImageNet-21k are as good or better than networks pre-trained on largest available medical X-Ray data when performing transfer to large X-Ray targets. We conclude that substantially increasing model and generic, medical domain-agnostic natural image source data scale in the pre-training can enable high quality out-of-domain transfer to medical domain specific targets, removing dependency on large medical domain-specific source data often not available in the practice., Comment: Short version published in MedNeurIPS 2021. Long version published in IJCNN 2022. Code: https://github.com/SLAMPAI/large-scale-pretraining-transfer

To test generalization ability of a class of deep neural networks, we randomly generate a large number of different rule sets for 2-D cellular automata (CA), based on John Conway's Game of Life. Using these rules, we compute several trajectories for each CA instance. A deep convolutional encoder-decoder network with short and long range skip connections is trained on various generated CA trajectories to predict the next CA state given its previous states. Results show that the network is able to learn the rules of various, complex cellular automata and generalize to unseen configurations. To some extent, the network shows generalization to rule sets and neighborhood sizes that were not seen during the training at all. Code to reproduce the experiments is publicly available at: https://github.com/SLAMPAI/generalization-cellular-automata, Comment: Accepted at 23rd International Conference on Artificial Intelligence (July 2021, Las Vegas, USA) To appear in: Springer Transactions on Computational Science & Computational Intelligence

We apply adversarial domain adaptation in unsupervised setting to reduce sample bias in a supervised high energy physics events classifier training. We make use of a neural network containing event and domain classifier with a gradient reversal layer to simultaneously enable signal versus background events classification on the one hand, while on the other hand minimising the difference in response of the network to background samples originating from different MC models via adversarial domain classification loss. We show the successful bias removal on the example of simulated events at the LHC with $t\bar{t}H$ signal versus $t\bar{t}b\bar{b}$ background classification and discuss implications and limitations of the method, Comment: 17 pages, 8 figures, to be submitted to MLST

The Obstacle Tower Challenge is the task to master a procedurally generated chain of levels that subsequently get harder to complete. Whereas the most top performing entries of last year's competition used human demonstrations or reward shaping to learn how to cope with the challenge, we present an approach that performed competitively (placed 7th) but starts completely from scratch by means of Deep Reinforcement Learning with a relatively simple feed-forward deep network structure. We especially look at the generalization performance of the taken approach concerning different seeds and various visual themes that have become available after the competition, and investigate where the agent fails and why. Note that our approach does not possess a short-term memory like employing recurrent hidden states. With this work, we hope to contribute to a better understanding of what is possible with a relatively simple, flexible solution that can be applied to learning in environments featuring complex 3D visual input where the abstract task structure itself is still fairly simple., Comment: 8 pages, 9 figures, 2 tables, under review

Turbulence is still one of the main challenges for accurately predicting reactive flows. Therefore, the development of new turbulence closures which can be applied to combustion problems is essential. Data-driven modeling has become very popular in many fields over the last years as large, often extensively labeled, datasets became available and training of large neural networks became possible on GPUs speeding up the learning process tremendously. However, the successful application of deep neural networks in fluid dynamics, for example for subgrid modeling in the context of large-eddy simulations (LESs), is still challenging. Reasons for this are the large amount of degrees of freedom in realistic flows, the high requirements with respect to accuracy and error robustness, as well as open questions, such as the generalization capability of trained neural networks in such high-dimensional, physics-constrained scenarios. This work presents a novel subgrid modeling approach based on a generative adversarial network (GAN), which is trained with unsupervised deep learning (DL) using adversarial and physics-informed losses. A two-step training method is used to improve the generalization capability, especially extrapolation, of the network. The novel approach gives good results in a priori as well as a posteriori tests with decaying turbulence including turbulent mixing. The applicability of the network in complex combustion scenarios is furthermore discussed by employing it to a reactive LES of the Spray A case defined by the Engine Combustion Network (ECN)., Comment: Submitted to Combustion Symposium 2020

Growing neuropsychological and neurophysiological evidence suggests that the visual cortex uses parts-based representations to encode, store and retrieve relevant objects. In such a scheme, objects are represented as a set of spatially distributed local features, or parts, arranged in stereotypical fashion. To encode the local appearance and to represent the relations between the constituent parts, there has to be an appropriate memory structure formed by previous experience with visual objects. Here, we propose a model how a hierarchical memory structure supporting efficient storage and rapid recall of parts-based representations can be established by an experience-driven process of self-organization. The process is based on the collaboration of slow bidirectional synaptic plasticity and homeostatic unit activity regulation, both running at the top of fast activity dynamics with winner-take-all character modulated by an oscillatory rhythm. These neural mechanisms lay down the basis for cooperation and competition between the distributed units and their synaptic connections. Choosing human face recognition as a test task, we show that, under the condition of open-ended, unsupervised incremental learning, the system is able to form memory traces for individual faces in a parts-based fashion. On a lower memory layer the synaptic structure is developed to represent local facial features and their interrelations, while the identities of different persons are captured explicitly on a higher layer. An additional property of the resulting representations is the sparseness of both the activity during the recall and the synaptic patterns comprising the memory traces., Comment: 34 pages, 12 Figures, 1 Table, published in Frontiers in Computational Neuroscience (Special Issue on Complex Systems Science and Brain Dynamics), http://www.frontiersin.org/neuroscience/computationalneuroscience/paper/10.3389/neuro.10/015.2009/