Oral
Static Automatic Batching In TensorFlow
Ashish Agarwal

Tue Jun 11th 05:00 -- 05:05 PM @ Grand Ballroom

Dynamic neural networks are becoming increasingly common, and yet it is hard to implement them efficiently. One-the-fly operation batching for such models is sub-optimal and suffers from run time overheads, while writing manually batched versions can be hard and error-prone. To address this we extend TensorFlow with pfor, a parallel-for loop optimized using static loop vectorization. With pfor, users can express computation using nested loops and conditional constructs, but get performance resembling that of a manually batched version. Benchmarks demonstrate speedups of one to two orders of magnitude on range of tasks, from jacobian computation, to TreeLSTMs.

Author Information

Ashish Agarwal (Google Brain)

Related Events (a corresponding poster, oral, or spotlight)