Filter by Track: All TracksUser TutorialDeveloper TutorialMain StageOpen Source LoungeEdge & Embedded LoungeLightning Talks
December 15 - TVM Tutorial Day
| Title | Time (PST) | Track | Speakers | Organization |
|---|---|---|---|---|
| TVM User Tutorials | 8:55 am | Main Stage | Speakers: | Organization: The Apache TVM Community |
| TVM Developer Tutorials | 8:55 am | Open Source Lounge | Speakers: | Organization: The Apache TVM Community |
| Introduction to TVM with TVMC: A Command-Line Driver for TVM | 9:00 am | User Tutorial | Speakers: Leandro Nunes Gustavo Romero | Organization: Arm, Linaro |
| TVM Developer Bootcamp – Overview | 9:00 am | Developer Tutorial | Speakers: Josh Fromm | Organization: OctoML |
| TVM Developer Bootcamp – Frontends | 9:30 am | Developer Tutorial | Speakers: Josh Fromm | Organization: OctoML |
| TVM Developer Bootcamp – Operators | 10:00 am | Developer Tutorial | Speakers: Josh Fromm | Organization: OctoML |
| Introduction to TVM with the Python/TVMC Interface | 10:30 am | User Tutorial | Speakers: Jocelyn Shiue | Organization: OctoML |
| TVM Developer Bootcamp – Writing Hardware Backends | 10:30 am | Developer Tutorial | Speakers: Eric Lunderberg | Organization: OctoML |
| Understanding TensorIR: An Abstraction for Tensorized Program Optimization | 11:00 am | User Tutorial | Speakers: Siyuan Feng | Organization: Shanghai Jiao Tong University |
| Neural Network Quantization with Brevitas | 11:15 am | Developer Tutorial | Speakers: Alessandro Pappalardo | Organization: Xilinx Research |
| Build on AWS for Startups Overview | 12:15 pm | Developer Tutorial | Speakers: AWS | Organization: AWS |
| Using Swift and Apache TVM to Develop ML Apps for the Apple Ecosystem | 12:45 pm | User Tutorial | Speakers: Iliya Slavutin | Organization: Deelvin Solutions |
| Optimizing ML Models for Arduino with microTVM | 1:00 pm | User Tutorial | Speakers: Gavin Uberti | Organization: OctoML |
| Automating and Simplifying MLPerf Inference Benchmark Submissions with TVM and CK | 2:00 pm | User Tutorial | Speakers: Grigori Fursin Thomas Zhu Alexander Peskov | Organization: OctoML |
December 16 - TVMCon Day 1
| Title | Time (PST) | Track | Speakers | Organization |
|---|---|---|---|---|
| Keynote and Community Update | 9:00 am | Main Stage | Speakers: Tianqi Chen Jared Roesch Denise Kutnick | Organization: OctoML |
| Growing from “Neat Project” to “Industry Leader” | 9:50 am | Main Stage | Speakers: Sarah Novotny | Organization: Microsoft Azure Office of CTO |
| How do we Responsibly Democratize Transformers, the Most Impactful Tech of the Decade? | 10:15 am | Main Stage | Speakers: Clément Delangue | Organization: Hugging Face |
| Break | 10:30 am | Main Stage | Speakers: | Organization: |
| TVM at Xilinx: Compiling for Versal AI Engines | 10:40 am | Main Stage | Speakers: Jorn Tuyls | Organization: Xilinx |
| TVM at Arm | 11:00 am | Main Stage | Speakers: Steve Roddy | Organization: Arm |
| Deep Learning Model Training at AWS | 11:20 am | Main Stage | Speakers: Yida Wang | Organization: Amazon Web Services |
| PyTorch AOTAutograd: A pythonic extension point for integrating compilers into PyTorch training | 11:40 am | Main Stage | Speakers: Horace He | Organization: Meta |
| Lunch Break | 12:00 pm | Main Stage | Speakers: | Organization: |
| Sponsor Booths Open | 12:00 pm | Open Source Lounge | Speakers: | Organization: TVMCon Sponsors |
| Machine Learning Compilers: Bringing ML to the Edge | 12:45 pm | Main Stage | Speakers: Chip Huyen | Organization: Stealth Startup and Stanford University |
| Miniaturizing Models for microNPUs: a Cascading Scheduler for TVM | 12:45 pm | Edge & Embedded Lounge | Speakers: Matthew Barrett | Organization: Arm |
| TVM Unity Community Session | 12:45 pm | Open Source Lounge | Speakers: Jared Roesch | Organization: OctoML |
| Lightning Quick Performance on Apple M1 with TVM | 1:10 pm | Main Stage | Speakers: Phil Mazenett | Organization: OctoML |
| Better Tensor Core Support in TVM with CUTLASS | 1:15 pm | Open Source Lounge | Speakers: Leyuan Wang Andrew Kerr | Organization: Bytedance |
| Unified Static Memory Planner (USMP) | 1:15 pm | Edge & Embedded Lounge | Speakers: Manupa Karunaratne Dmitriy Smirnov | Organization: Arm |
| TVM in the Arene AI Platform of Woven Planet | 1:20 pm | Main Stage | Speakers: Ryo Takahashi | Organization: Woven Planet Holdings, Inc. |
| Transparent TVM Backend Acceleration | 1:35 pm | Main Stage | Speakers: Tiejun Chen | Organization: VMWare |
| Integrating the Synopsys DSP-enhanced ARC Processors into TVM | 1:40 pm | Edge & Embedded Lounge | Speakers: Kuen-Wey Lin David Wang Chuck Pilkington | Organization: Industrial Technology Research Institute, Synopsys |
| Common Subexpression Elimination for TIR | 1:45 pm | Open Source Lounge | Speakers: Franck Slama | Organization: Qualcomm |
| TVM Streamer – Accelerating Multimedia Framework with TVM | 1:50 pm | Main Stage | Speakers: Cecilia Albertsson Hiroki Endoh Shinya Kaji | Organization: Fixstars Corporation, NTT TechnoCross Corporation |
| TVM with PaddlePaddle | 2:05 pm | Main Stage | Speakers: Jiajun Jiang Yuguang Deng Yin Ma | Organization: Baidu |
| RAZOR: Relay via PyTorch Lazy Tensor | 2:15 pm | Open Source Lounge | Speakers: Haozheng Fan | Organization: Amazon |
| TVM at Marvell | 2:20 pm | Main Stage | Speakers: Derek Chickles Chien-Chun (Joe) Chou | Organization: Marvell, Machine Learning Software |
| Break | 2:30 pm | Main Stage | Speakers: | Organization: |
| DietCode: Automatic Optimization for Dynamic Tensor Programs | 2:30 pm | Open Source Lounge | Speakers: Bojian Zheng Ziheng Jiang | Organization: University of Toronto |
| Lightning Talks – Compilation and Optimization | 2:40 pm | Lightning Talks | Speakers: Various | Organization: The Apache TVM Community |
| Puffin: A MLIR-based tensorizer for training | 2:40 pm | Lightning Talks | Speakers: Liangfu Chen | Organization: Amazon Web Services |
| Verifiable Compilation to Accelerators Using a Formal Software/Hardware Interface and Flexible Matching | 2:46 pm | Lightning Talks | Speakers: Gus Smith | Organization: University of Washington |
| Automatic Tensorized Program Optimization | 2:52 pm | Lightning Talks | Speakers: Bohan Hou | Organization: Carnegie Mellon University |
| Tsunami: Training in TVM with Relay | 2:58 pm | Lightning Talks | Speakers: Altan Haan | Organization: OctoML |
| R-TVM, a polyhedral mapper for TVM | 3:04 pm | Lightning Talks | Speakers: Benoit Meister | Organization: Reservoir Labs |
| Fuzzing TVM Relay | 3:10 pm | Lightning Talks | Speakers: Steven Lyubomirsky | Organization: University of Washington |
| RoofTune:Accelerating the Tuning Process Based on Roofline Model | 3:16 pm | Lightning Talks | Speakers: Hui Zhong Haiwen Fu Xiaohua Shi | Organization: Beihang University |
| SONAR: Direct architecture and system optimization search | 3:22 pm | Lightning Talks | Speakers: Elias Jaasaari | Organization: Carnegie Mellon University |
| Collage: Automated Integration of Deep Learning Backends | 3:28 pm | Lightning Talks | Speakers: Byungsoo Jeon | Organization: Carnegie Mellon University |
| The CoRa Tensor Compiler: Compilation for Ragged Tensors with Minimal Padding | 3:34 pm | Lightning Talks | Speakers: Pratik Fegade | Organization: Carnegie Mellon University |
| Multi-stream Support for Virtual Machine Executor | 3:40 pm | Lightning Talks | Speakers: Yaoyao Ding | Organization: University of Toronto |
December 17 - TVMCon Day 2
| Title | Time (PST) | Track | Speakers | Organization |
|---|---|---|---|---|
| Keynote and Community Update | 9:00 am | Main Stage | Speakers: Luis Ceze Jason Knight Vanessa Yan Sameer Farooqui | Organization: OctoML |
| Green AI | 9:50 am | Main Stage | Speakers: Jesse Dodge Oren Etzioni | Organization: Allen Institute for AI |
| Monitoring ML on the Edge | 10:10 am | Main Stage | Speakers: Alessya Visnjic | Organization: WhyLabs |
| Break | 10:30 am | Main Stage | Speakers: | Organization: |
| Nexus: A TVM-based Inference Server | 10:40 am | Main Stage | Speakers: Loc Huynh | Organization: Microsoft |
| The Torch-MLIR Project | 10:55 am | Main Stage | Speakers: Sean Silva | Organization: Google |
| TVM For Snapdragon Mobile Platform | 11:10 am | Main Stage | Speakers: Jeff Gehlhaar | Organization: Qualcomm |
| Efficient Application Mapping in a Heterogeneous Environment using TVM | 11:25 am | Main Stage | Speakers: Joey Chou Alicja Kwasniewska and Sharath Raghava | Organization: SiMa.ai |
| Seamless Cross Platform Deployment for Communities | 11:40 am | Main Stage | Speakers: Wilson Yu Mei Ye Amir Khojaste Galeshkhale | Organization: AMD |
| Lunch Break | 12:00 pm | Main Stage | Speakers: | Organization: |
| Sponsor Booths Open | 12:00 pm | Open Source Lounge | Speakers: | Organization: TVMCon Sponsors |
| Improvements to TVM Developer Productivity | 12:45 pm | Open Source Lounge | Speakers: Andrew Reusch | Organization: OctoML |
| TVM at Tencent | 12:45 pm | Main Stage | Speakers: Wenxi Zhu | Organization: Tencent |
| Reflections on microTVM on Cortex-M | 12:45 pm | Edge & Embedded Lounge | Speakers: Chris Sidebottom | Organization: Arm |
| TPAT: TensorRT Plugin Autogen Tool | 1:05 pm | Main Stage | Speakers: Qian Qiu | Organization: Tencent |
| Lightweight Profiling for TVM | 1:15 pm | Open Source Lounge | Speakers: Jyotsna Verma | Organization: Qualcomm |
| How to Use FP16 Quantization in TVM | 1:15 pm | Edge & Embedded Lounge | Speakers: Andrew Luo | Organization: OctoML |
| ONNX, ONNX Runtime, and TVM | 1:20 pm | Main Stage | Speakers: Manash Goswami Ganesan Ramalingam | Organization: Microsoft |
| TVM at EdgeCortix – Heterogenous AI Hardware Acceleration | 1:35 pm | Main Stage | Speakers: Antonio Tomas | Organization: EdgeCortix Inc. |
| TVM-Enabled Realtime Character Control for Mobile Games | 1:45 pm | Open Source Lounge | Speakers: Haidong Lan | Organization: Tencent |
| VTA++: Expanded Design Space Exploration with an Enhanced Versatile Tensor Accelerator | 1:45 pm | Edge & Embedded Lounge | Speakers: Abhijit Davare | Organization: Intel Labs |
| Relax: Co-Designing High-Level Abstraction Towards TVM Unity | 1:50 pm | Main Stage | Speakers: Yuchen Jin Altan Haan Ziheng Jiang | Organization: OctoML |
| microTVM in 2021 | 2:10 pm | Main Stage | Speakers: Andrew Reusch Mehrdad Hessar | Organization: OctoML |
| Apple BNNS Integration, Benefits and Challenges | 2:15 pm | Open Source Lounge | Speakers: Alexander Peskov | Organization: Deelvin Solutions |
| Break | 2:30 pm | Main Stage | Speakers: | Organization: |
| Lightning Talks – Edge and Embedded | 2:40 pm | Lightning Talks | Speakers: Various | Organization: Apache TVM Community |
| Extremely Fast GEMM on AVX512 CPUs Combining TVM and XSMM | 2:40 pm | Lightning Talks | Speakers: Honglin Zhu | Organization: Tencent Holding Limited |
| A Simulation-Based TVM Extension for Auto-Tuning in the Early Design Phases | 2:46 pm | Lightning Talks | Speakers: Yannick Braatz Michael J. Klaiber Axel J. Acosta | Organization: Robert Bosch GmbH |
| Leveraging TVM as a front-end compiler for RISC-V based custom tinyML processor | 2:52 pm | Lightning Talks | Speakers: Vaibhav Verma | Organization: University of Virginia |
| Integrate TVM With MediaTek NeuroPilot for Mobile Devices | 2:58 pm | Lightning Talks | Speakers: Robert Lai Chun-Ping Chung | Organization: Mediatek |
| Caffe-SSD Inference on Edge Device Using TVM and Hybrid Script | 3:04 pm | Lightning Talks | Speakers: Masahiro Hiramori | Organization: Mitsubishi Electric Corporation |
| Support sparse and irregular workloads in TIR | 3:10 pm | Lightning Talks | Speakers: Zihao Ye Ruihang Lai | Organization: University of Washington |
| Optimization with TVM Hybrid OP on RISC-V with P Extension | 3:16 pm | Lightning Talks | Speakers: Lee Jenq-Kuen Yu Meng-Shiun Yuan Chuan-Yue | Organization: National Tsing Hua University, Taiwan |
| Integer Only Quantization and Pruning of Transformers | 3:22 pm | Lightning Talks | Speakers: Amir Gholami Sehoon Kim | Organization: UC Berkeley |
| Another Automatic Kernel Generator on Huawei NPU with Polyhedral Compilation | 3:28 pm | Lightning Talks | Speakers: Shenghu Jiang Mingli Sun | Organization: Beihang University |
| Extending auto-scheduler to support performance evaluation with timing model | 3:34 pm | Lightning Talks | Speakers: Minchun Liao Yaohua Chen Chungta King | Organization: National Tsing Hua University, Industrial Technology Research Institute (ITRI) |
| TVM at Qualcomm Adreno | 3:40 pm | Lightning Talks | Speakers: Siva Rama Krishna | Organization: Qualcomm India Pvt Ltd |