top of page

How Fashable Revolutionizes Fashion Design with Azure Machine Learning and PyTorch


Fashable, a Portugal-based company, is dedicated to revolutionizing the fashion industry by harnessing the power of AI technology. Within the fashion sector, significant challenges such as unsustainable manufacturing practices, surplus unsold inventory, and protracted production cycles persist. Nevertheless, Orlando Ribas Fernandes, the Co-Founder and CEO of Fashable, is confident that technology can effectively mitigate these issues. Through the utilization of Azure Machine Learning and PyTorch, Fashable has engineered an AI algorithm with the capacity to generate distinctive clothing designs. This groundbreaking advancement empowers fashion enterprises to more effectively meet consumer demands, expedite time-to-market processes, and reduce clothing wastage.


PyTorch and Azure Machine Learning are the perfect match for our research team goals, saving time to create disruptive innovation.

Orlando Ribas Fernandes: Co-Founder and CEO

Fashable



A style explosion


Fashion is adored worldwide, being one of the most rapidly expanding, profitable, and demanding industries. It sets high standards for swift turnaround times, innovative designs, and a continuous flow of fresh styles.


What is driving this surge in production? According to some experts, the phenomenon known as "fast fashion" is to blame, referring to the mass production of runway trends in a quick and cost-effective manner. While luxury designers may spend months (or even years) creating a collection, fast fashion labels can accomplish this in a fraction of the time and, thanks to low-cost labor and materials, at a significantly lower price.


How Fashable Revolutionizes Fashion Design with Azure Machine Learning and PyTorch


An unstable trajectory


While fast fashion benefits customers by offering designer looks at lower prices, critics argue that it contributes to the emergence of a disposable culture. Some studies suggest that the average consumer disposes of a piece of clothing after just seven to ten wears.


The increase in clothing production combined with shorter lifecycles has resulted in a growing landfill issue. Annually, the United States alone discards 21.6 billion pounds of textile waste. In a world that is increasingly focused on sustainability, fashion designers must find ways to meet demand while minimizing waste.



Bringing outside changes to an insider industry


Orlando Ribas Fernandes, Co-Founder and CEO of Fashable, is an AI expert with a background in technology and 15 years of experience in artificial intelligence research and development. He collaborated with Microsoft to create innovative AI technologies for fashion designers to engage directly with customers, showcasing the industry's potential for positive change.


The Fashable AI app can swiftly generate numerous unique clothing designs using AI, all within minutes and without the need for physical materials. The next objective was to identify a suitable platform to support this capability. "As we delved into our AI technology, one of our key criteria was to have a platform and deep learning framework that resonated with our vision and strategy. We sought a platform equipped with the necessary tools to push the boundaries of innovation. This requirement led us to consider options such as PyTorch and Microsoft Azure Machine Learning," explained Ribas Fernandes.


With PyTorch, Fashable benefits from an open-source machine learning framework known for its power, flexibility, and speed, supported by a thriving developer community. PyTorch on Microsoft Azure streamlines deep-learning tasks, eliminating the need for complex setups. Azure Machine Learning offers extensive GPU support and seamless integration, enabling Fashable to train image-based models without costly investments. By leveraging low-priority virtual machines, Fashable has significantly reduced training costs within Azure Machine Learning studio. The combination of Azure Machine Learning and PyTorch enhances Fashable's machine learning projects with efficient MLOps capabilities.


chart on how fashable's machine learning work

Fashable recognized the benefits of leveraging a Distributed Data-Parallel implementation with Azure Container for PyTorch (ACPT). According to Ribas Fernandes, the organization was able to achieve effective and efficient model execution across multi-modal nodes, resulting in successful outcomes that were unattainable in alternative environments.



Sustainable fast fashion


Creating a new fashion collection demands a substantial investment of time, finances, and resources. The ever-changing landscape of fashion trends presents a formidable obstacle for designers in accurately forecasting market demand for their creations.


Fashable simplifies the design process and reduces uncertainty. Its advanced AI technology leverages a range of neural networks to analyze information from diverse channels like social media and e-commerce platforms to grasp current trends, styles, and garment preferences. These sophisticated models continually adjust to the dynamic fashion environment, facilitating real-time visual enhancements to digital designs, such as altering sleeve lengths or transitioning patterns from stripes to polka dots.


user interface of fashable's and how AI generate product images work

Designers can use Fashable to A/B test creations on social media, enabling quick interest and demand forecasts before production. The platform streamlines the process, reducing the time from design to department store from months to minutes. This allows designers to innovate, market directly to customers, and forecast demand efficiently.



An intelligent future


While Fashable primarily focuses on leveraging its technology within the fashion industry, Ribas Fernandes has indicated that the company is diversifying its applications into other sectors such as healthcare, manufacturing, and the metaverse.


Specifically within healthcare, Fashable's AI technology can play a significant role in tasks like detecting lung cancer. Through the analysis of x-ray scans of both healthy and diseased lungs, Fashable's model is trained to identify signs of cancer and other medical conditions. Ribas Fernandes expresses optimism regarding the potential of AI to improve the accuracy and quality of medical diagnoses, ultimately facilitating timely treatment for patients.


Ribas Fernandes envisions a bright future for the fusion of artificial intelligence and human creativity. "I am deeply passionate about AI and firmly believe in its ethical application to not only save lives but also to revolutionize our world."



1 view0 comments
bottom of page