diff --git a/_community_stories/1.md b/_community_stories/1.md new file mode 100644 index 000000000000..267bd361773f --- /dev/null +++ b/_community_stories/1.md @@ -0,0 +1,7 @@ +--- +title: 'How Outreach Productionizes PyTorch-based Hugging Face Transformers for NLP' +ext_url: https://www.databricks.com/blog/2021/05/14/how-outreach-productionizes-pytorch-based-hugging-face-transformers-for-nlp.html +date: May 14, 2021 +tags: ["Advertising & Marketing"] +--- +At Outreach, a leading sales engagement platform, our data science team is a driving force behind our innovative product portfolio largely driven by deep learning and AI. We recently announced enhancements to the Outreach Insights feature, which is powered by the proprietary Buyer Sentiment deep learning model developed by the Outreach Data Science team. This model allows sales teams to deepen their understanding of customer sentiment through the analysis of email reply content, moving from just counting the reply rate to classification of the replier’s intent. \ No newline at end of file diff --git a/_community_stories/10.md b/_community_stories/10.md new file mode 100644 index 000000000000..b7ee0b245571 --- /dev/null +++ b/_community_stories/10.md @@ -0,0 +1,7 @@ +--- +title: 'Solliance makes headlines with cryptocurrency news analysis platform powered by Azure Machine Learning, PyTorch' +ext_url: https://medium.com/pytorch/solliance-makes-headlines-with-cryptocurrency-news-analysis-platform-powered-by-azure-machine-52a2a290fefb +date: Mar 14, 2022 +tags: ["Finance"] +--- +Solliance delivers cutting-edge solutions that fill gaps across a wide variety of industries. Through its recent collaboration with Baseline, Solliance revolutionizes the cryptocurrency trading experience, extracting news insights from more than 150,000 global sources in near real time. To manage Baseline workloads, Solliance brought Microsoft Azure Machine Learning and PyTorch together for maximum processing power and deep learning capabilities. The result: investors can get under the headlines and see which specific news metrics are moving the volatile crypto market to make more informed trading decisions, while Baseline can release new features in weeks instead of months. \ No newline at end of file diff --git a/_community_stories/11.md b/_community_stories/11.md new file mode 100644 index 000000000000..96138278f774 --- /dev/null +++ b/_community_stories/11.md @@ -0,0 +1,7 @@ +--- +title: 'Create a Wine Recommender Using NLP on AWS' +ext_url: https://www.capitalone.com/tech/machine-learning/create-wine-recommender-using-nlp/ +date: March 2, 2022 +tags: ["Finance"] +--- +In this tutorial, we’ll build a simple machine learning pipeline using a BERT word embedding model and the Nearest Neighbor algorithm to recommend wines based on user inputted preferences. To create and power this recommendation engine, we’ll leverage AWS’s SageMaker platform, which provides a fully managed way for us to train and deploy our service. \ No newline at end of file diff --git a/_community_stories/12.md b/_community_stories/12.md new file mode 100644 index 000000000000..56f6b2ab93ed --- /dev/null +++ b/_community_stories/12.md @@ -0,0 +1,7 @@ +--- +title: 'Crayon boosts speed, accuracy of healthcare auditing process using Azure Machine Learning and PyTorch' +ext_url: https://www.microsoft.com/en/customers/story/1503427278296945327-crayon-partner-professional-services-azure +date: June 28, 2022 +tags: ["Healthcare"] +--- +Healthcare providers need to be able to verify that they’re maintaining the highest operating safety and efficacy standards. Those standards are set by a national accreditation organization whose surveyors, often healthcare professionals themselves, regularly visit facilities and document situations that might need to be corrected or brought back in line with the latest rules and policies. That assessment and accreditation process generates a huge amount of data, and even the most experienced surveyors struggle to keep ahead of the ongoing development of thousands of policy rules that might be relevant in any particular scenario. Vaagan and his team took on the task of fixing the issue by building a machine learning solution that could ingest text from those reports and return a top ten list of the latest associated rules with unprecedented accuracy. They used Azure technology, development tools, and services to bring that solution to fruition. Crayon customers report clear time savings with the new healthcare solution. Just as important, the solution provides consistent responses that aren’t subject to the vagaries of individual interpretation or potentially out-of-date data. \ No newline at end of file diff --git a/_community_stories/13.md b/_community_stories/13.md new file mode 100644 index 000000000000..0e7b6371eaf1 --- /dev/null +++ b/_community_stories/13.md @@ -0,0 +1,7 @@ +--- +title: 'Extracting value from siloed healthcare data using federated learning with Azure Machine Learning' +ext_url: https://www.microsoft.com/en/customers/story/1587521717158304168-microsoft-partner-professional-services-azure +date: December 30, 2022 +tags: ["Healthcare"] +--- +Sensitive information such as healthcare data is often siloed within health organization boundaries. This has posed a challenge to machine learning models used by the health and life sciences industry that require data for training purposes. To improve patient care and accelerate health industry progression, the Microsoft Health & Life Sciences AI group used a federated learning setup to train their biomedical natural language processing service, Text Analytics for Health, while preserving the trust boundaries of siloed data. The federated learning framework was built using Microsoft Azure Machine Learning and open-source technologies to help organizations analyze siloed data and build new applications without compromising data privacy. \ No newline at end of file diff --git a/_community_stories/14.md b/_community_stories/14.md new file mode 100644 index 000000000000..23f3a2bbc3f8 --- /dev/null +++ b/_community_stories/14.md @@ -0,0 +1,7 @@ +--- +title: 'HippoScreen Improves AI Performance by 2.4x with oneAPI Tools' +ext_url: https://www.intel.com/content/www/us/en/developer/articles/case-study/hipposcreen-boosts-ai-performance-2-4x-with-oneapi.html +date: Feb 21, 2023 +tags: ["Healthcare"] +--- +The Taiwan-based neurotechnology startup used tools and frameworks in the Intel® oneAPI Base and AI Analytics Toolkits to the improve efficiency and build times of deep-learning models used in its Brain Waves AI system. As a result, HippoScreen is able to broaden the system’s applications to a wider range of psychiatric conditions and diseases. \ No newline at end of file diff --git a/_community_stories/16.md b/_community_stories/16.md new file mode 100644 index 000000000000..0bee1f4ac29a --- /dev/null +++ b/_community_stories/16.md @@ -0,0 +1,7 @@ +--- +title: "Disney's Creative Genome by Miquel Farré" +ext_url: https://www.youtube.com/watch?v=KuDxEhHk2Rk +date: Apr 27, 2021 +tags: ["Media & Entertainment"] +--- +Miquel Farré is a senior technology manager at Disney, taking the lead on projects at the intersection of video technology, machine learning and web applications. Metadata that drives content searchability is most often indexed at the title level, with limited governance and high ambiguity; at best, keyword metadata has been added to a title as a layer of enrichment. \ No newline at end of file diff --git a/_community_stories/17.md b/_community_stories/17.md new file mode 100644 index 000000000000..3669cda5942f --- /dev/null +++ b/_community_stories/17.md @@ -0,0 +1,7 @@ +--- +title: 'How Disney uses PyTorch for animated character recognition' +ext_url: https://medium.com/pytorch/how-disney-uses-pytorch-for-animated-character-recognition-a1722a182627 +date: Jul 16, 2020 +tags: ["Media & Entertainment"] +--- +The long and incremental evolution of the media industry, from a traditional broadcast and home video model, to a more mixed model with increasingly digitally-accessible content, has accelerated the use of machine learning and artificial intelligence (AI). Advancing the implementation of these technologies is critical for a company like Disney that has produced nearly a century of content, as it allows for new consumer experiences and enables new applications for illustrators and writers to create the highest-quality content. \ No newline at end of file diff --git a/_community_stories/18.md b/_community_stories/18.md new file mode 100644 index 000000000000..87dc0045b4ec --- /dev/null +++ b/_community_stories/18.md @@ -0,0 +1,7 @@ +--- +title: 'Machine Learning at Tubi: Powering Free Movies, TV and News for All' +ext_url: https://medium.com/pytorch/machine-learning-at-tubi-powering-free-movies-tv-and-news-for-all-51499643018e +date: Feb 25, 2021 +tags: ["Media & Entertainment"] +--- +In this blog series, our aim is to highlight the nuances of Machine Learning in Tubi’s Ad-based Video on Demand (AVOD) space as practiced at Tubi. Machine Learning helps solve myriad problems involving recommendations, content understanding and ads. We extensively use PyTorch for several of these use cases as it provides us the flexibility, computational speed and ease of implementation to train large scale deep neural networks using GPUs. \ No newline at end of file diff --git a/_community_stories/19.md b/_community_stories/19.md new file mode 100644 index 000000000000..1c26fc2f71a2 --- /dev/null +++ b/_community_stories/19.md @@ -0,0 +1,7 @@ +--- +title: 'How Pixar uses AI and GANs to create high-resolution content' +ext_url: https://venturebeat.com/business/how-pixar-uses-ai-and-gans-to-create-high-resolution-content/ +date: July 17, 2020 +tags: ["Media & Entertainment"] +--- +As digital animators continue to push the boundaries of technology and creativity, the technical teams that support them are turning to artificial intelligence and machine learning to deliver the tools they need. That’s the case at Pixar, where the company has made new machine learning breakthroughs it hopes will both improve quality and reduce costs. \ No newline at end of file diff --git a/_community_stories/2.md b/_community_stories/2.md new file mode 100644 index 000000000000..424e66e6fcac --- /dev/null +++ b/_community_stories/2.md @@ -0,0 +1,7 @@ +--- +title: 'Amazon Ads Uses PyTorch and AWS Inferentia to Scale Models for Ads Processing' +ext_url: /blog/amazon-ads-case-study/ +date: February 24, 2022 +tags: ["Advertising & Marketing", "Retail"] +--- +Amazon Ads uses PyTorch, TorchServe, and AWS Inferentia to reduce inference costs by 71% and drive scale out. Amazon Ads helps companies build their brand and connect with shoppers through ads shown both within and beyond Amazon’s store, including websites, apps, and streaming TV content in more than 15 countries. Businesses and brands of all sizes, including registered sellers, vendors, book vendors, Kindle Direct Publishing (KDP) authors, app developers, and agencies can upload their own ad creatives, which can include images, video, audio, and, of course, products sold on Amazon. \ No newline at end of file diff --git a/_community_stories/20.md b/_community_stories/20.md new file mode 100644 index 000000000000..c5ad56b5e728 --- /dev/null +++ b/_community_stories/20.md @@ -0,0 +1,7 @@ +--- +title: 'Running BERT model inference on AWS Inf1: From model compilation to speed comparison' +ext_url: https://note.com/asahi_ictrad/n/nf5195eb53b88 +date: November 21, 2021 +tags: ["Media & Entertainment"] +--- +In this tech blog, we will compare the speed and cost of Inferentia, GPU, and CPU for a BERT sequence labeling example. We also provide a helpful tutorial on the steps for model compilation and inference on Inf1 instances. \ No newline at end of file diff --git a/_community_stories/21.md b/_community_stories/21.md new file mode 100644 index 000000000000..ede721b4241e --- /dev/null +++ b/_community_stories/21.md @@ -0,0 +1,7 @@ +--- +title: 'Ambient Clinical Intelligence: Generating Medical Reports with PyTorch' +ext_url: /blog/ambient-clinical-intelligence-generating-medical-reports-with-pytorch/ +date: May 12, 2022 +tags: ["Medical"] +--- +Complete and accurate clinical documentation is an essential tool for tracking patient care. It allows for treatment plans to be shared among care teams to aid in continuity of care and ensures a transparent and effective process for reimbursement. \ No newline at end of file diff --git a/_community_stories/22.md b/_community_stories/22.md new file mode 100644 index 000000000000..24683262ecfd --- /dev/null +++ b/_community_stories/22.md @@ -0,0 +1,7 @@ +--- +title: 'AstraZeneca is using PyTorch-powered algorithms to discover new drugs' +ext_url: https://www.zdnet.com/article/astrazeneca-is-using-pytorch-powered-algorithms-to-discover-new-drugs/ +date: Sept. 30, 2020 +tags: ["Medical"] +--- +Since it launched in 2017, Facebook's machine-learning framework PyTorch has been put to good use, with applications ranging from powering Elon Musk's autonomous cars to driving robot-farming projects. Now pharmaceutical firm AstraZeneca has revealed how its in-house team of engineers are tapping PyTorch too, and for equally as important endeavors: to simplify and speed up drug discovery. \ No newline at end of file diff --git a/_community_stories/23.md b/_community_stories/23.md new file mode 100644 index 000000000000..ffda0ce4b314 --- /dev/null +++ b/_community_stories/23.md @@ -0,0 +1,7 @@ +--- +title: 'Deploying huggingface‘s BERT to production with pytorch/serve' +ext_url: https://medium.com/analytics-vidhya/deploy-huggingface-s-bert-to-production-with-pytorch-serve-27b068026d18 +date: Apr 25, 2020 +tags: ["Medical"] +--- +TL;DR: pytorch/serve is a new awesome framework to serve torch models in production. This story teaches you how to use it for huggingface/transformers models like BERT. \ No newline at end of file diff --git a/_community_stories/24.md b/_community_stories/24.md new file mode 100644 index 000000000000..fb33da259dd6 --- /dev/null +++ b/_community_stories/24.md @@ -0,0 +1,7 @@ +--- +title: 'How AI is Helping Vets to Help our Pets' +ext_url: https://medium.com/pytorch/how-ai-is-helping-vets-to-help-our-pets-e6e3d58c052e +date: Sep 7, 2021 +tags: ["Medical"] +--- +1 in 4 dogs, and 1 in 5 cats, will develop cancer at some point in their lives. Pets today have a better chance of being successfully treated than ever, thanks to advances in early recognition, diagnosis and treatment. \ No newline at end of file diff --git a/_community_stories/25.md b/_community_stories/25.md new file mode 100644 index 000000000000..5b2905604d25 --- /dev/null +++ b/_community_stories/25.md @@ -0,0 +1,7 @@ +--- +title: 'How theator Built a Continuous Training Framework To Scale up Its Surgical Intelligence Platform' +ext_url: https://medium.com/pytorch/how-theator-built-a-continuous-training-framework-to-scale-up-its-surgical-intelligence-platform-b5135e3229fd +date: Dec 17, 2020 +tags: ["Medical"] +--- +Performing surgery is largely about decision making. As Dr. Frank Spencer put it in 1978, “A skillfully performed operation is about 75% decision making and 25% dexterity”. Five decades later, and the surgical field is finally — albeit gradually — implementing advances in data science and AI to enhance surgeons’ ability to make the best decisions in the operating room. That’s where theator comes in: the company is re-imagining surgery with a Surgical Intelligence platform that leverages highly advanced AI, specifically machine learning and computer vision technology, to analyze every step, event, milestone, and critical junction of surgical procedures — significantly boosting surgeons’ overall performance. \ No newline at end of file diff --git a/_community_stories/26.md b/_community_stories/26.md new file mode 100644 index 000000000000..63397a1af6dc --- /dev/null +++ b/_community_stories/26.md @@ -0,0 +1,7 @@ +--- +title: 'Speeding up drug discovery with advanced machine learning' +ext_url: https://medium.com/pytorch/speeding-up-drug-discovery-with-advanced-machine-learning-b17d59e0daa6 +date: Sep 30, 2020 +tags: ["Medical"] +--- +Whatever our job title happens to be at AstraZeneca, we’re seekers. I’m part of the Biological Insights Knowledge Graph (BIKG) team. We help scientists comb through massive amounts of data in our quest to find the information we need to help us deliver life-changing medicines. \ No newline at end of file diff --git a/_community_stories/27.md b/_community_stories/27.md new file mode 100644 index 000000000000..d612e75e5724 --- /dev/null +++ b/_community_stories/27.md @@ -0,0 +1,7 @@ +--- +title: 'Using PyTorch to streamline machine-learning projects' +ext_url: https://www.zdnet.com/article/using-pytorch-to-streamline-machine-learning-projects/ +date: Dec. 17, 2020 +tags: ["Medical"] +--- +For many surgeons, the possibility of going back into the operating room to review the actions they carried out on a patient could provide invaluable medical insights. \ No newline at end of file diff --git a/_community_stories/28.md b/_community_stories/28.md new file mode 100644 index 000000000000..a77212f18930 --- /dev/null +++ b/_community_stories/28.md @@ -0,0 +1,7 @@ +--- +title: 'Run inference at scale for OpenFold, a PyTorch-based protein folding ML model, using Amazon EKS' +ext_url: https://aws.amazon.com/blogs/machine-learning/run-inference-at-scale-for-openfold-a-pytorch-based-protein-folding-ml-model-using-amazon-eks/ +date: Oct. 25, 2022 +tags: ["Medical"] +--- +In drug discovery, understanding the 3D structure of proteins is key to assessing the ability of a drug to bind to it, directly impacting its efficacy. Predicting the 3D protein form, however, is very complex, challenging, expensive, and time consuming, and can take years when using traditional methods such as X-ray diffraction. Applying machine learning (ML) to predict these structures can significantly accelerate the time to predict protein structures—from years to hours. Several high-profile research teams have released algorithms such as AlphaFold2 (AF2), RoseTTAFold, and others. These algorithms were recognized by Science magazine as the 2021 Breakthrough of the Year. \ No newline at end of file diff --git a/_community_stories/29.md b/_community_stories/29.md new file mode 100644 index 000000000000..a6ac02477809 --- /dev/null +++ b/_community_stories/29.md @@ -0,0 +1,7 @@ +--- +title: 'Optimize Protein Folding Costs with OpenFold on AWS Batch' +ext_url: https://aws.amazon.com/blogs/hpc/optimize-protein-folding-costs-with-openfold-on-aws-batch/ +date: Oct. 4, 2022 +tags: ["Medical"] +--- +Knowing the physical structure of proteins is an important part of the drug discovery process. Machine learning (ML) algorithms like AlphaFold v2.0 significantly reduce the cost and time needed to generate usable protein structures. These projects have also inspired development of AI-driven workflows for de novo protein design and protein-ligand interaction analysis. \ No newline at end of file diff --git a/_community_stories/3.md b/_community_stories/3.md new file mode 100644 index 000000000000..99394598af83 --- /dev/null +++ b/_community_stories/3.md @@ -0,0 +1,9 @@ +--- +title: 'NASA and IBM to Speed AI Creation with New Foundation Models' +ext_url: https://thenewstack.io/nasa-and-ibm-to-speed-ai-creation-with-new-foundation-models/ +date: February 2, 2023 +tags: ["Aerospace"] +--- +NASA and IBM are working together to create foundation models based on NASA’s data sets — including geospatial data — with the goal of accelerating the creation of AI models. + +Foundation models are trained on large, broad data sets, then used to train other AI models by using targeted and smaller datasets. Foundation models can be used for different tasks and can apply information about one situation to another. One real-world example of a foundation model at work is ChatGPT3, which was built with the foundation model, GPT3. \ No newline at end of file diff --git a/_community_stories/30.md b/_community_stories/30.md new file mode 100644 index 000000000000..1a723fb9bc9a --- /dev/null +++ b/_community_stories/30.md @@ -0,0 +1,7 @@ +--- +title: 'How Datarock is using PyTorch for more intelligent mining decision making' +ext_url: https://medium.com/pytorch/how-datarock-is-using-pytorch-for-more-intelligent-decision-making-d5d1694ba170 +date: Jun 9, 2020 +tags: ["Mining"] +--- +The mining industry is currently going through a digital revolution as it looks for new and innovative ways to explore and extract mineral resources. This has largely been driven by a need to reduce costs in a competitive global industry that’s experiencing declining ore grades and fewer new discoveries. \ No newline at end of file diff --git a/_community_stories/32.md b/_community_stories/32.md new file mode 100644 index 000000000000..b58f986c159a --- /dev/null +++ b/_community_stories/32.md @@ -0,0 +1,7 @@ +--- +title: 'How Trigo built a scalable AI development & deployment pipeline for Frictionless Retail' +ext_url: https://medium.com/pytorch/how-trigo-built-a-scalable-ai-development-deployment-pipeline-for-frictionless-retail-b583d25d0dd +date: Jun 16, 2020 +tags: ["Retail"] +--- +Trigo is a provider of AI & computer vision based checkout-free systems for the retail market, enabling frictionless checkout and a range of other in-store operational and marketing solutions such as predictive inventory management, security and fraud prevention, pricing optimization and event-driven marketing. \ No newline at end of file diff --git a/_community_stories/33.md b/_community_stories/33.md new file mode 100644 index 000000000000..423906b5bc15 --- /dev/null +++ b/_community_stories/33.md @@ -0,0 +1,7 @@ +--- +title: 'How We Built: An Early-Stage Recommender System' +ext_url: https://www.onepeloton.com/press/articles/designing-an-early-stage-recommender-system +date: October 18, 2021 +tags: ["Retail"] +--- +Personalization is ubiquitous on most platforms today. Supercharged by connectivity, and scaled by machine learning, most experiences on the internet are tailored to our personal tastes. Peloton classes offer a diversity of instructors, languages, fitness disciplines, durations and intensity. Each Member has specific fitness goals, schedule, fitness equipment, and level of skill or strength. This diversity of content and individuality of Member needs at massive scale creates the opportunity for a recommender system to create a personalized experience on the Peloton platform. \ No newline at end of file diff --git a/_community_stories/34.md b/_community_stories/34.md new file mode 100644 index 000000000000..8fc6ba0c4738 --- /dev/null +++ b/_community_stories/34.md @@ -0,0 +1,7 @@ +--- +title: 'Automated Background Removal in E-commerce Fashion Image Processing Using PyTorch on Databricks' +ext_url: https://www.databricks.com/blog/2021/04/29/automated-background-removal-in-e-commerce-fashion-image-processing-using-pytorch-on-databricks.html +date: April 29, 2021 +tags: ["Retail"] +--- +Wehkamp is one of the biggest e-commerce companies in the Netherlands, with more than 500,000 daily visitors on their website. A wide variety of products offered on the Wehkamp site aims to meet its customers’ many needs. An important aspect of any customer visit on an e-commerce website is a qualitative and accurate visual experience of the products. At a large scale, this is no easy task, with thousands of product photos processed in a local photo studio. \ No newline at end of file diff --git a/_community_stories/35.md b/_community_stories/35.md new file mode 100644 index 000000000000..c572513c77ea --- /dev/null +++ b/_community_stories/35.md @@ -0,0 +1,8 @@ +--- +title: 'Search Model Serving Using PyTorch and TorchServe' +ext_url: https://medium.com/walmartglobaltech/search-model-serving-using-pytorch-and-torchserve-6caf9d1c5f4d +date: Jan 23, 2023 +tags: ["Retail"] +--- +Walmart Search has embarked on the journey of adopting Deep Learning in the search ecosystem to improve search relevance. For our pilot use case, we served the computationally intensive Bert Base model at runtime with an objective to achieve low latency and high throughput. + diff --git a/_community_stories/36.md b/_community_stories/36.md new file mode 100644 index 000000000000..5a2b3e7c9737 --- /dev/null +++ b/_community_stories/36.md @@ -0,0 +1,8 @@ +--- +title: 'How We Used AWS Inferentia to Boost PyTorch NLP Model Performance by 4.9x for the Autodesk Ava Chatbot' +ext_url: https://medium.com/pytorch/how-we-used-aws-inferentia-to-boost-pytorch-nlp-model-performance-by-4-9x-9f79f5314ca8 +date: Apr 7, 2021 +tags: ["Technology"] +--- +Autodesk is a multinational software company with world-renowned products in areas such as Architecture, Engineering, & Construction, Manufacturing, and Media & Entertainment. Amongst Autodesk’s best-known products are AutoCAD, Revit, Maya, and Fusion 360. The company has millions of customers around the world, and many of them have need for support to make best use of their products. + diff --git a/_community_stories/37.md b/_community_stories/37.md new file mode 100644 index 000000000000..a7e6e376a9e0 --- /dev/null +++ b/_community_stories/37.md @@ -0,0 +1,7 @@ +--- +title: 'Bentley Systems creates breakthrough framework, drastically speeds up AI development with Azure Machine Learning' +ext_url: https://www.microsoft.com/en/customers/story/1480221307332639219-bentley-systems-partner-professional-services-azure-machine-learning +date: March 16, 2022 +tags: ["Technology"] +--- +Software innovator Bentley Systems offers a broad portfolio of solutions to help the organizations that design, build, and operate the world’s infrastructure assets. The company uses machine learning in its flagship product to read disparate paper-based asset data and transform it into consolidated digital data. To speed up and formalize this process, Bentley created a machine learning operations framework using Microsoft Azure Machine Learning and PyTorch. Developers’ speed and job satisfaction have shot up since they began using this stable, reproducible framework, which easily gets their code into the cloud, accelerating delivery by three to five times and significantly increasing efficiency. \ No newline at end of file diff --git a/_community_stories/38.md b/_community_stories/38.md new file mode 100644 index 000000000000..e76ae4a1164e --- /dev/null +++ b/_community_stories/38.md @@ -0,0 +1,7 @@ +--- +title: 'PyTorch Community Voices' +ext_url: https://www.youtube.com/watch?v=LBOIxA5sg2A +date: Jun 2, 2021 +tags: ["Technology"] +--- +Join us for an interview with star PyTorch community members Alexander O’Connor and Binghui Ouyang from AutoDesk as we learn how they used PyTorch and AWS Inferentia to deploy production-scale models in chatbot intent classification. \ No newline at end of file diff --git a/_community_stories/39.md b/_community_stories/39.md new file mode 100644 index 000000000000..d7771ef6c0a6 --- /dev/null +++ b/_community_stories/39.md @@ -0,0 +1,7 @@ +--- +title: 'How PyTorch is bringing the power of AI to computers and smartphones' +ext_url: https://ai.meta.com/blog/pytorch-ai-smartphones-computers/ +date: December 2, 2022 +tags: ["Technology"] +--- +Many of the experiences people enjoy on Facebook and Instagram are powered by artificial intelligence (AI). A number of them, like Assistant, Avatars, and AR effects, cannot be powered by server-side AI due to latency, network bandwidth, and other constraints. Running AI on-device —that is, directly on a phone, tablet, or even a pair of smart glasses — offers huge advantages over constantly sending data back to a server. It’s faster, and it creates a privacy-enhancing experience for people who use our platforms. However, on-device AI presents new challenges, since it requires coping with devices that have a small battery, far less powerful processors, and less memory than a server in a data center. \ No newline at end of file diff --git a/_community_stories/4.md b/_community_stories/4.md new file mode 100644 index 000000000000..90f2c15de2ec --- /dev/null +++ b/_community_stories/4.md @@ -0,0 +1,8 @@ +--- +title: 'AI for AG: Production machine learning for agriculture' +ext_url: https://medium.com/pytorch/ai-for-ag-production-machine-learning-for-agriculture-e8cfdb9849a1 +date: Aug 6, 2020 +tags: ["Agriculture"] +--- +How did farming affect your day today? If you live in a city, you might feel disconnected from the farms and fields that produce your food. Agriculture is a core piece of our lives, but we often take it for granted. + diff --git a/_community_stories/40.md b/_community_stories/40.md new file mode 100644 index 000000000000..0c45ff732658 --- /dev/null +++ b/_community_stories/40.md @@ -0,0 +1,7 @@ +--- +title: 'Axon offers technology boost for public safety with in-car Automated License Plate Recognition on Azure' +ext_url: https://www.microsoft.com/en/customers/story/1610624764549732009-axon-partner-professional-services-azure +date: March 09, 2023 +tags: ["Technology"] +--- +Axon, a technology leader in public safety, developed AI technology to add cutting-edge license plate recognition capabilities to its in-car camera products, which now can identify plates for vehicles of interest and provide law enforcement with proactive notifications and alerts. Axon AI scientists and engineers chose Microsoft Azure infrastructure as a scalable, cost-efficient, and feature-rich environment where they can develop and test AI models. With Azure compute, storage, and PyTorch and machine learning resources, Axon can easily take advantage of the latest software and hardware technology to develop best-in-class AI solutions for its customers. \ No newline at end of file diff --git a/_community_stories/41.md b/_community_stories/41.md new file mode 100644 index 000000000000..bd1d083e7577 --- /dev/null +++ b/_community_stories/41.md @@ -0,0 +1,7 @@ +--- +title: 'ML Model Server Resource Saving - Transition From High-Cost GPUs to Intel CPUs and oneAPI powered Software with performance' +ext_url: /blog/ml-model-server-resource-saving/ +date: October 11, 2023 +tags: ["Technology"] +--- +Here, We will be sharing our experience in moving AI workloads from our GPU servers to our Intel CPU servers without any performance or quality degradation, and saving annual costs of approximately 340 thousand U.S. Dollar (refer to the Conclusion) in the process. \ No newline at end of file diff --git a/_community_stories/42.md b/_community_stories/42.md new file mode 100644 index 000000000000..21fb9616f644 --- /dev/null +++ b/_community_stories/42.md @@ -0,0 +1,7 @@ +--- +title: 'Dialogue Assistance for Customer Service at Airbnb' +ext_url: https://www.youtube.com/watch?v=jtVUV0Gzxp0&t=730s +date: Aug 20, 2019 +tags: ["Technology"] +--- +Businesses are using PyTorch, an open source machine learning framework, to seamlessly build, train, and deploy AI models in production across their products and services. Hear how industry leaders leverage PyTorch to help power everything from ubiquitous productivity software used across the world to enabling advances in medicine for fighting cancer. \ No newline at end of file diff --git a/_community_stories/43.md b/_community_stories/43.md new file mode 100644 index 000000000000..a51d7765b881 --- /dev/null +++ b/_community_stories/43.md @@ -0,0 +1,7 @@ +--- +title: 'Using deep learning and PyTorch to power next gen aircraft at Caltech' +ext_url: https://www.youtube.com/watch?v=se206WBk2dM +date: Nov 14, 2019 +tags: ["Research", "Aeorospace"] +--- +Learn how Caltech’s Center for Autonomous Systems and Technologies (CAST) uses PyTorch to build deep learning systems that can understand the aerodynamics of how aircrafts interact with the ground to enable much smoother and safer landings. \ No newline at end of file diff --git a/_community_stories/44.md b/_community_stories/44.md new file mode 100644 index 000000000000..4ab96977bba0 --- /dev/null +++ b/_community_stories/44.md @@ -0,0 +1,7 @@ +--- +title: 'Deepset achieves a 3.9x speedup and 12.8x cost reduction for training NLP models by working with AWS and NVIDIA' +ext_url: https://aws.amazon.com/blogs/machine-learning/deepset-achieves-a-3-9x-speedup-and-12-8x-cost-reduction-for-training-nlp-models-by-working-with-aws-and-nvidia/ +date: Jan 27, 2021 +tags: ["Research", "NLP"] +--- +At deepset, we’re building the next-level search engine for business documents. Our core product, Haystack, is an open-source framework that enables developers to utilize the latest NLP models for semantic search and question answering at scale. Our software as a service (SaaS) platform, Haystack Hub, is used by developers from various industries, including finance, legal, and automotive, to find answers in all kinds of text documents. You can use these answers to improve the search experience, cover the long-tail of chat bot queries, extract structured data from documents, or automate invoicing processes. \ No newline at end of file diff --git a/_community_stories/45.md b/_community_stories/45.md new file mode 100644 index 000000000000..6ad0704a27e1 --- /dev/null +++ b/_community_stories/45.md @@ -0,0 +1,7 @@ +--- +title: 'PyTorch at Dolby Labs' +ext_url: https://www.youtube.com/watch?v=K5hD0et_wUc&list=PL_lsbAsL_o2BY-RrqVDKDcywKnuUTp-f3&index=20 +date: Nov 6, 2019 +tags: ["Research", "NLP"] +--- +Hear how Dolby Labs is using PyTorch to develop deep learning for audio, and learn about the challenges that audio AI presents and the breakthroughs and applications they’ve built at Dolby to push the field forward. \ No newline at end of file diff --git a/_community_stories/46.md b/_community_stories/46.md new file mode 100644 index 000000000000..d7562ccc49bb --- /dev/null +++ b/_community_stories/46.md @@ -0,0 +1,7 @@ +--- +title: 'Using a Grapheme to Phoneme Model in Cisco’s Webex Assistant' +ext_url: https://blogs.cisco.com/developer/graphemephoneme01 +date: September 7, 2021 +tags: ["Research", "NLP"] +--- +Grapheme to Phoneme (G2P) is a function that generates pronunciations (phonemes) for words based on their written form (graphemes). It has an important role in automatic speech recognition systems, natural language processing, and text-to-speech engines. In Cisco’s Webex Assistant, we use G2P modelling to assist in resolving person names from voice. See here for further details of various techniques we use to build robust voice assistants. \ No newline at end of file diff --git a/_community_stories/47.md b/_community_stories/47.md new file mode 100644 index 000000000000..c479e32d0c4d --- /dev/null +++ b/_community_stories/47.md @@ -0,0 +1,7 @@ +--- +title: 'AI21 Labs Trains 178-Billion-Parameter Language Model Using Amazon EC2 P4d Instances, PyTorch' +ext_url: https://aws.amazon.com/solutions/case-studies/AI21-case-study-p4d/ +date: June 7, 2021 +tags: ["Research", "NLP"] +--- +AI21 Labs uses machine learning to develop language models focused on understanding meaning, and in 2021 it set a goal to train the recently released Jurassic-1 Jumbo, an autoregressive language model with 178 billion parameters. Developers who register for beta testing will get access to Jurassic-1 Jumbo and can immediately start to customize the model for their use case. The software startup wanted to train the model efficiently, so it looked to Amazon Web Services (AWS) and built a solution using Amazon Elastic Compute Cloud (Amazon EC2), a web service that provides secure, resizable compute capacity in the cloud. Choosing Amazon EC2 gave the company control over the training process, including node allocation. \ No newline at end of file diff --git a/_community_stories/48.md b/_community_stories/48.md new file mode 100644 index 000000000000..147c55460932 --- /dev/null +++ b/_community_stories/48.md @@ -0,0 +1,7 @@ +--- +title: 'The Why and How of Scaling Large Language Models' +ext_url: https://www.youtube.com/watch?v=qscouq3lo0s +date: Jan 4, 2022 +tags: ["Research", "NLP"] +--- +Anthropic is an AI safety and research company that’s working to build reliable, interpretable, and steerable AI systems. Over the past decade, the amount of compute used for the largest training runs has increased at an exponential pace. We've also seen in many domains that larger models are able to attain better performance following precise scaling laws. The compute needed to train these models can only be attained using many coordinated machines that are communicating data between them. In this talk, Nicholas Joseph (Technical Staff, Anthropic) goes through why and how they can scale up training runs to use these machines efficiently. \ No newline at end of file diff --git a/_community_stories/49.md b/_community_stories/49.md new file mode 100644 index 000000000000..8dac0320ec2f --- /dev/null +++ b/_community_stories/49.md @@ -0,0 +1,7 @@ +--- +title: 'University of Pécs enables text and speech processing in Hungarian, builds the BERT-large model with just 1,000 euro with Azure' +ext_url: https://www.microsoft.com/en/customers/story/1402696956382669362-university-of-pecs-higher-education-azure-en-hungary +date: August 10, 2021 +tags: ["Research", "NLP"] +--- +Everyone prefers to use their mother tongue when communicating with chat agents and other automated services. However, for languages like Hungarian—spoken by only 15 million people—the market size will often be viewed as too small for large companies to create software, tools or applications that can process Hungarian text as input. Recognizing this need, the Applied Data Science and Artificial Intelligence team from University of Pécs decided to step up. Using Microsoft AI Solutions and ONNX Runtime solutions, it built and trained its own BERT-large model in native Hungarian in under 200 hours and total build cost of 1,000 euro. \ No newline at end of file diff --git a/_community_stories/5.md b/_community_stories/5.md new file mode 100644 index 000000000000..b0006022eece --- /dev/null +++ b/_community_stories/5.md @@ -0,0 +1,7 @@ +--- +title: 'Using PyTorch for Monocular Depth Estimation Webinar' +ext_url: https://www.youtube.com/watch?v=xf2QgioY370 +date: Sep 27, 2024 +tags: ["Research"] +--- +In this webinar, Bob Chesebrough of Intel guides you through the steps he took to create a clipped image with background clutter removed from the image. He accomplished this using monocular depth estimation with PyTorch. This could potentially be used to automate structure from motion and other image-related tasks where you want to highlight or focus on a single portion of an image, particularly for identifying parts of the image that were closest to the camera. Specifically, he used depth estimation on a couple of images that he took at a natural history museum to capture just the dinosaur in the foreground, eliminating the background murals, lights, and building structure. The cool thing about this algorithm is that it creates a depth estimate from a single image! \ No newline at end of file diff --git a/_community_stories/50.md b/_community_stories/50.md new file mode 100644 index 000000000000..9f1014e46b5d --- /dev/null +++ b/_community_stories/50.md @@ -0,0 +1,7 @@ +--- +title: 'Mapillary Research: Seamless Scene Segmentation and In-Place Activated BatchNorm' +ext_url: /blog/mapillary-research/ +date: July 23, 2019 +tags: ["Research"] +--- +With roads in developed countries like the US changing up to 15% annually, Mapillary addresses a growing demand for keeping maps updated by combining images from any camera into a 3D visualization of the world. Mapillary’s independent and collaborative approach enables anyone to collect, share, and use street-level images for improving maps, developing cities, and advancing the automotive industry. \ No newline at end of file diff --git a/_community_stories/51.md b/_community_stories/51.md new file mode 100644 index 000000000000..2b9e820aa47a --- /dev/null +++ b/_community_stories/51.md @@ -0,0 +1,7 @@ +--- +title: 'How 3DFY.ai Built a Multi-Cloud, Distributed Training Platform Over Spot Instances with TorchElastic and Kubernetes' +ext_url: https://medium.com/pytorch/how-3dfy-ai-built-a-multi-cloud-distributed-training-platform-over-spot-instances-with-44be40936361 +date: Jun 17, 2021 +tags: ["Research"] +--- +Deep Learning development is becoming more and more about minimizing the time from idea to trained model. To shorten this lead time, researchers need access to a training environment that supports running multiple experiments concurrently, each utilizing several GPUs. \ No newline at end of file diff --git a/_community_stories/52.md b/_community_stories/52.md new file mode 100644 index 000000000000..4d249134c9ea --- /dev/null +++ b/_community_stories/52.md @@ -0,0 +1,7 @@ +--- +title: 'SearchSage: Learning Search Query Representations at Pinterest' +ext_url: https://medium.com/pinterest-engineering/searchsage-learning-search-query-representations-at-pinterest-654f2bb887fc +date: Nov 9, 2021 +tags: ["Research"] +--- +Pinterest surfaces billions of ideas to people every day, and the neural modeling of embeddings for content, users, and search queries are key in the constant improvement of these machine learning-powered recommendations. Good embeddings — representations of discrete entities as vectors of numbers — enable fast candidate generation and are strong signals to models that classify, retrieve and rank relevant content. \ No newline at end of file diff --git a/_community_stories/53.md b/_community_stories/53.md new file mode 100644 index 000000000000..7929cd8495db --- /dev/null +++ b/_community_stories/53.md @@ -0,0 +1,7 @@ +--- +title: 'IBM Research: Bringing massive AI models to any cloud' +ext_url: https://research.ibm.com/blog/ibm-pytorch-cloud-ai-ethernet +date: Nov 17, 2022 +tags: ["Research"] +--- +The field of AI is in the middle of a revolution. In recent years, AI models have made images, songs, or even websites out of simple text prompts. These types of models with billions of parameters, called foundation models, can with little fine-tuning be repurposed from one task to another, removing countless hours of training and labelling, and refitting a model to take on a new task. \ No newline at end of file diff --git a/_community_stories/54.md b/_community_stories/54.md new file mode 100644 index 000000000000..a6e2e0b4a958 --- /dev/null +++ b/_community_stories/54.md @@ -0,0 +1,7 @@ +--- +title: 'ChemicalX: A Deep Learning Library for Drug Pair Scoring' +ext_url: https://arxiv.org/abs/2202.05240 +date: Feb 10, 2022 +tags: ["Research", "Healthcare"] +--- +In this paper, we introduce ChemicalX, a PyTorch-based deep learning library designed for providing a range of state of the art models to solve the drug pair scoring task. The primary objective of the library is to make deep drug pair scoring models accessible to machine learning researchers and practitioners in a streamlined this http URL design of ChemicalX reuses existing high level model training utilities, geometric deep learning, and deep chemistry layers from the PyTorch ecosystem. Our system provides neural network layers, custom pair scoring architectures, data loaders, and batch iterators for end users. We showcase these features with example code snippets and case studies to highlight the characteristics of ChemicalX. A range of experiments on real world drug-drug interaction, polypharmacy side effect, and combination synergy prediction tasks demonstrate that the models available in ChemicalX are effective at solving the pair scoring task. Finally, we show that ChemicalX could be used to train and score machine learning models on large drug pair datasets with hundreds of thousands of compounds on commodity hardware. \ No newline at end of file diff --git a/_community_stories/55.md b/_community_stories/55.md new file mode 100644 index 000000000000..103aa76737c0 --- /dev/null +++ b/_community_stories/55.md @@ -0,0 +1,7 @@ +--- +title: 'Graph Convolutional Operators in the PyTorch JIT' +ext_url: https://www.youtube.com/watch?v=4swsvOLzL_A&list=PL_lsbAsL_o2BSe3eS4spnodObBa3RL08E&index=3 +date: Dec 2, 2020 +tags: ["Research", "Science"] +--- +In this talk, scientist Lindsey Gray and Ph.D. student Matthias Fey co-examine how the challenges of High Energy Particle Physics are driving the need for more efficient research and development pipelines in neural network development. In particular, they look at the additions made to PyTorch Geometric, which allow Graph Neural Network models to be compiled by the PyTorch JIT, significantly easing the process of deploying such networks at scale. \ No newline at end of file diff --git a/_community_stories/6.md b/_community_stories/6.md new file mode 100644 index 000000000000..b218ca839725 --- /dev/null +++ b/_community_stories/6.md @@ -0,0 +1,7 @@ +--- +title: 'How Wadhwani AI Uses PyTorch To Empower Cotton Farmers' +ext_url: https://medium.com/pytorch/how-wadhwani-ai-uses-pytorch-to-empower-cotton-farmers-14397f4c9f2b +date: Oct 22, 2020 +tags: ["Agriculture"] +--- +Cotton is a major fibre crop across the world, cultivated in over 80 countries with nearly 100 million families across the world rely on cotton farming for their livelihood. With such importance placed on many farmers’ crops, cotton’s particular vulnerability to pest infestations has been troubling to many. However, pest infestation is also simultaneously one of the most significant and preventable problems that farmers face with 55% of all pesticide usage in India being devoted to cotton farming. \ No newline at end of file diff --git a/_community_stories/7.md b/_community_stories/7.md new file mode 100644 index 000000000000..7103bf45be6c --- /dev/null +++ b/_community_stories/7.md @@ -0,0 +1,7 @@ +--- +title: 'How Lyft Uses PyTorch to Power Machine Learning for Their Self-Driving Cars' +ext_url: https://medium.com/pytorch/how-lyft-uses-pytorch-to-power-machine-learning-for-their-self-driving-cars-80642bc2d0ae +date: Oct 7, 2020 +tags: ["Autonomous Driving"] +--- +Lyft’s mission is to improve people’s lives with the world’s best transportation. We believe in a future where self-driving cars make transportation safer and more accessible for everyone. That’s why Level 5, Lyft’s self-driving division, is developing a complete autonomous system for the Lyft network to provide riders’ access to the benefits of this technology. However, this is an incredibly complex task. \ No newline at end of file diff --git a/_community_stories/8.md b/_community_stories/8.md new file mode 100644 index 000000000000..f23672204a07 --- /dev/null +++ b/_community_stories/8.md @@ -0,0 +1,7 @@ +--- +title: 'Wayve’s AV2.0 builds a brighter future with Azure Machine Learning and PyTorch' +ext_url: https://www.microsoft.com/en/customers/story/1415185921593450824-wayve-partner-professional-services-azure-machine-learning +date: May 25, 2022 +tags: ["Autonomous Driving"] +--- +Wayve wants to accelerate and scale autonomous vehicle (AV) development by using vision-based machine learning for rapid prototyping and quick iteration. So, it developed a platform that uses the open-source machine learning framework PyTorch with Microsoft Azure Machine Learning to gather, manage, and process millions of hours of driving data per year—petabytes of data—consisting of images, GPS data, and data from other sensors. Wayve now has the scalable capacity to build and iterate driving models for complex urban environments, adjust models more nimbly, and adapt to new environments more readily. \ No newline at end of file diff --git a/_community_stories/9.md b/_community_stories/9.md new file mode 100644 index 000000000000..0d208d53d26d --- /dev/null +++ b/_community_stories/9.md @@ -0,0 +1,8 @@ +--- +title: 'AI Helps Duolingo Personalize Language Learning' +ext_url: https://aws.amazon.com/machine-learning/customers/innovators/duolingo/ +date: May 25, 2024 +tags: ["Education"] +--- +Learning a foreign language was probably one of your goals last year. And the year before, and the year before that. Like gym memberships, our best intentions often don’t survive very long. Aside from the time required to achieve proficiency with a new language, most people struggle with traditional approaches to learning. Even many web-based language tools can be monotonous and cumbersome. + diff --git a/community-stories.html b/community-stories.html index ded3c547b2ca..84f0e2395229 100644 --- a/community-stories.html +++ b/community-stories.html @@ -2,51 +2,47 @@ layout: default title: Community Stories permalink: /community-stories -body-class: comm-stories +body-class: blog background-class: comm-stories-background ---

Community Stories

-

Learn how our community solves real, everyday machine learning problems with PyTorch

+

Read case studies on how our community solves real, everyday machine learning problems with PyTorch

-
-
-
-
-
- -
+
+
+
+
+ + {% assign community_stories = site.community_stories | sort_natural: "date" | reverse %} + {% for post in community_stories %} +
+
+

{{ post.date | date: '%B %d, %Y' }}

+

+ {{ post.title }} +

+

+ {% if post.tags %} + Industry: + {% for tag in post.tags %} + {{ tag }}{% unless forloop.last %}, {% endunless %} + {% endfor %} + {% endif %} +

+

{{ post.excerpt | strip_html | truncate: 500}}

+
+ + Read More + +
+ {% endfor %} +
- {% include production.html %} - {% include research.html %}
- -