The machine learning industry has continued its dramatic growth over the past year as businesses sped up their plans for automation and digitization. The need to run machine learning models quickly and efficiently is as strong as ever. But with new hardware chips, software frameworks and optimization techniques being released at a lightning pace, how do you make sense of all this progress?
TVMcon is the global event for the AI community to share and learn about the latest advancements in machine learning acceleration. This year’s virtual conference will bring together researchers, developers and users in the ML acceleration space to collaborate on what happens next in the state of the art of deep learning compilation. And the conference will be free for everyone to attend!
Day one will offer tutorials to get beginners and experts in machine learning up to speed with TVM. Over the next two days, the main event will feature a full schedule of interactive sessions covering recent advances in ML frameworks, compilers, systems and architecture support, and business case studies.
What is Apache TVM?
Apache TVM enables access to high-performance machine learning anywhere for everyone. As TVM's diverse community of hardware vendors, compiler engineers and ML researchers work together to build a unified, programmable software stack, they enrich the entire ML technology ecosystem, making it accessible to the wider ML community. TVM empowers users to leverage community-driven ML-based optimizations to amplify the reach of their research and development, which in turn raises the collective performance of all ML while driving its cost down.
TVM accomplishes this by providing a new accelerating compiler that exposes graph-level and operator-level optimizations targeting a variety of platforms, covering the entire spectrum of server-class GPUs, mobile phones, embedded devices, and new accelerators. Since its initial pull request, TVM has grown steadily, with more than 7,000 contributions from over 600 contributors across more than 40 organizations. TVM has demonstrated the value of solving the problem of how to accelerate machine learning models, regardless of the platform on which they were trained or the hardware where they will be deployed.
What started as an academic research project has grown into a community focused on pushing the envelope of machine learning acceleration.
Call for proposals
TVM is part of a larger open-source effort dedicated to pushing the limits of machine learning to deploy models anywhere. We welcome papers from across the ML acceleration community, and are thrilled to share new work and advancements with the machine learning ecosystem. Previous years have included talks from leaders in the MLIR TensorFlow communities.
We’re looking for fresh ideas, unique perspectives and thought-provoking discussions that will give conference attendees insights into how to navigate a rapidly advancing landscape. We’d love to hear from you and encourage you to submit your work in ML acceleration for consideration.
We're excited to be bringing you TVMcon for a fourth year running, and are looking forward to welcoming you all to this free virtual event.