Abstract | ||
---|---|---|
Frameworks for writing, compiling, and optimizing deep learning (DL) models have recently enabled progress in areas like computer vision and natural language processing. Extending these frameworks to accommodate the rapidly diversifying landscape of DL models and hardware platforms presents challenging tradeoffs between expressiveness, composability, and portability. We present Relay, a new intermediate representation (IR) and compiler framework for DL models. The functional, statically-typed Relay IR unifies and generalizes existing DL IRs and can express state-of-the-art models. Relayu0027s expressive IR required careful design of the type system, automatic differentiation, and optimizations. Relayu0027s extensible compiler can eliminate abstraction overhead and target new hardware platforms. The design insights from Relay can be applied to existing frameworks to develop IRs that support extension without compromising on expressivity, composibility, and portability. Our evaluation demonstrates that the Relay prototype can already provide competitive performance for a broad class of models running on CPUs, GPUs, and FPGAs. |
Year | Venue | DocType |
---|---|---|
2019 | arXiv: Learning | Journal |
Volume | Citations | PageRank |
abs/1904.08368 | 0 | 0.34 |
References | Authors | |
0 | 9 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jared Roesch | 1 | 51 | 5.90 |
Steven Lyubomirsky | 2 | 0 | 0.34 |
Marisa Kirisame | 3 | 8 | 1.74 |
Josh Pollock | 4 | 0 | 0.34 |
Logan Weber | 5 | 0 | 0.34 |
Ziheng Jiang | 6 | 67 | 7.19 |
Tianqi Chen | 7 | 1887 | 83.63 |
Thierry Moreau | 8 | 105 | 8.54 |
Zachary Tatlock | 9 | 307 | 21.68 |