Application performance demands have increasingly been outpacing Moore’s law in a variety of fields, particularly AI and deep learning. Co-processors like GPUs offer immense speedup to applications in fields like AI And deep learning, compared to CPUs. At Bitfusion, we build technology to disaggregate co-processors like GPUs and re-aggregate them in real-time over Ethernet, Infiniband RDMA or RoCE network, to create an elastic AI infrastructure. Just like network attached storage, our technology allows customers to do network attached co-processors.
In 2018, I am excited to share with you my plans for the year ahead. The employees, investors, advisors, partners and customers have supported the growth of the company since it's founding, and I m very grateful for that. When I started this entrepreneurial journey three years ago to go after a big opportunity, heterogeneous hardware like GPUs, FPGAs for general purpose compute were just on the horizon and we set out on a mission to build the Operating System for the modern heterogeneous compute datacenter. We are now the world's first infrastructure management platform to support CPUs, GPUs and FPGAs, with AI as a beachhead. The team at Bitfusion has helped form a strong foundation for the company. We hope to accelerate and lead the transformation to elastic heterogeneous infrastructure deployment in the year ahead.
Deep learning and AI technologies are revolutionizing the world, whether it’s through self-driving cars, drones, virtual assistants, more accurate medical diagnosis, or automatic lead generation. As a result, AI is drastically altering the ways in which business is conducted. In the 90’s and 2000s, the web revolutionized businesses by offering them ability to improve customer value by an order of magnitude. Take Amazon for example which created a website for selling books which was brick and mortar until then, then later transformed the retail industry and the computing industry. Another example is how Mobile revolutionized businesses just under a decade ago.
Over the last 10 or so years, application performance demands have increasingly been outpacing Moore’s law in a variety of fields, particularly deep learning and AI. The solution has been to adopt heterogeneous accelerated processors, such as GPUs, FPGAs and various specialized ASICs. With the implementation of these alternative compute architectures, hardware has inevitably become more complex, and software more abstract, to keep up with the shifting landscape.
I recently read an article from the Huffington Post titled: “ Prospecting vs Retargeting: Making the Most of the Marketing Mix” . This article sparked this blog post because as a consumer I become infuriated when I shop online and I purchase an item to later get ads for the same thing I just purchased a couple of days ago. Clearly I don’t need another travel backpack because I just bought one. But hey, why not market some hiking shoes, or sleeping bags or ANYTHING other than the travel backpack I just bought. This is where AI and big data come into play.
Authors: Bhavesh Patel – Dell EMC & Mazhar Memon - Bitfusion Deep Learning (DL), a key technique driving artificial intelligence innovation, such as image recognition, chatbots, and self-driving cars, requires algorithms be ‘trained’ using large data sets. Initially, this can be done on a single node (server). However, as the models and datasets grow ever larger and more complex, it becomes essential to scale-out.
Over the past ten years, we have been noticing the trend of application performance demands starting to outpace Moore’s law in a variety of fields. The solution has been to rely on specialized processors like GPUs, FPGAs and other specialized ASICs. With these alternative compute architectures, hardware was becoming more complex and software was becoming more abstract.
Intro to TensorBoard Now that we’re constantly validating the data and saving our model, we can start thinking of ways to visualize the ins and outs of our model or ways to do exploratory data analysis of our model while or after it is done training. In Dandelion Mane’s talk at the TensorFlow Dev Summit 2017, he described it as a flashlight to shine on the black box of deep neural networks. Sometimes shining a bright light is ill-advised.
In our last post we gave a basic introduction to TensorFlow 1.0. What we want to do now is take our foundation and move it forward. One of the most important parts of deep learning is understanding what is going on while the code is running. As our problems get more complicated and our datasets get larger, training time can go from minutes to days. If we’ve picked a model with poor hyper-parameters or just a bad model in general, we don’t want to have to wait hours to make an adjustment to our model. Or if we have great hyperparameters and models, but don’t tell the model to train for enough steps we don't want to start from scratch. Or do we…
In a series of blog posts, we want to show a step-by-step guide on how to get from a basic TensorFlow model to best-in-class architectures. Deep learning is a hot topic and it’s easy to find starting and advanced resources, but it can be difficult to see how to get from the intro material to more advanced models. We at Bitfusion would like to help address the lack of points along the journey to becoming a Deep Learning expert.