The Apache TVM Community and OctoML closed out 2021 with the fourth annual Apache TVM and Open Source ML Acceleration Conference. It was the TVM community’s largest event ever, with 700 attendees from 34 nations coming together for a virtual conference featuring over 70 tutorials, sessions, lightning talks, and keynotes across 3 days. We want to thank everyone who made this event possible, including our speakers and sponsors, but most importantly the Apache TVM community. Whether you’re a long-time contributor or a first-time attendee, you’re part of a community that’s helping to build the ML stack for performance and portability, which enables access to high-performance machine learning anywhere for everyone.
If you missed the event, or want to revisit a talk, we have you covered. You can find all of the sessions from the conference on YouTube on the TVMCon 2021 Playlist If you’re looking for a specific session, there are direct links to every talk from the TVMCon schedule. Some of the conference highlights include:
The Day 1 Keynote from Jared Roesch, Tianqi Chen, and Denise Kutnick describing TVM Unity a new vision for the Apache TVM project that strengthens the virtuous circle connecting the ecosystem of ML scientists, engineers and hardware vendors.
The Day 2 Keynotefollowed up with Luis Ceze, CEO of OctoML, discussing the importance of automated ML acceleration in a rapidly changing software and hardware world. The presentation also includes a real-world demonstration of the OctoML platform, showing the end-to-end workflow of how models can be accelerated 2x (and upward) against a business as usual baseline–across a myriad of hardware choices–and then deployed in the cloud.
Several case studies of TVM in production from companies like Woven Planet
An overview of how TVM delivered Lightning Quick Performance on the Apple M1 chipset just weeks after its public release.
Thanks to everyone who made this the best TVMCon yet! You can learn more about the community over at the Apache TVM home page and Twitter feed, and you can find out more about how you can take advantage of TVM’s capabilities with an OctoML Platform Free Trial We look forward to seeing you in the community in the coming year!
The Future of Hardware is Software
General-purpose GPU computing helped launch the deep learning era. As ML models have grown larger and more computationally intense, however, they have changed the way GPUs are designed—and they have inspired a wave of new hardware that looks radically different from GPUs.
Flexible Systems for the Next ML Revolution(s)
The environment for machine learning innovation has never been better. Modern GPUs are marvels of supercomputer engineering. And software stacks have raised the level of abstraction for ML implementations...