Blake A. Hechtman
25 works · 2,041 citations · h-index 15
Top works
- Gemini: A Family of Highly Capable Multimodal Models
- Scaling Local Self-Attention for Parameter Efficient Visual Backbones
- Scaling Language Models: Methods, Analysis & Insights from Training Gopher
- Mesh-TensorFlow: Deep Learning for Supercomputers
- Heterogeneous-race-free memory models
- QuickRelease: A throughput-oriented approach to release consistency on GPUs
- Scale MLPerf-0.6 models on Google TPU-v3 Pods
- Overlap Communication with Dependent Computation via Decomposition in Large Deep Learning Models
- GSPMD: General and Scalable Parallelization for ML Computation Graphs
- Exploring memory consistency for massively-threaded throughput-oriented processors
- Large-Scale Discrete Fourier Transform on TPUs
- Evaluating cache coherent shared virtual memory for heterogeneous multicore chips
- Automatic Cross-Replica Sharding of Weight Update in Data-Parallel Training
- Unified Scaling Laws for Routed Language Models
- A Flexible Approach to Autotuning Multi-Pass Machine Learning Compilers
- Exploring the limits of Concurrency in ML Training on Google TPUs
- Heterogeneous-race-free memory models
- Sequential Consistency for Heterogeneous-Race-Free: Programmer-centric Memory Models for Heterogeneous Platforms
- Exploring memory consistency for massively-threaded throughput-oriented processors
- Heterogeneous-race-free memory models