Low Latency Transformer Inference on FPGAs for Physics Applications with hls4ml
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908045941932032 |
|---|---|
| author | Jiang, Zhixing Yin, Dennis Chen, Yihui Khoda, Elham E Hauck, Scott Hsu, Shih-Chieh Govorkova, Ekaterina Harris, Philip Loncar, Vladimir Moreno, Eric A. |
| author_facet | Jiang, Zhixing Yin, Dennis Chen, Yihui Khoda, Elham E Hauck, Scott Hsu, Shih-Chieh Govorkova, Ekaterina Harris, Philip Loncar, Vladimir Moreno, Eric A. |
| contents | This study presents an efficient implementation of transformer architectures in Field-Programmable Gate Arrays(FPGAs) using hls4ml. We demonstrate the strategy for implementing the multi-head attention, softmax, and normalization layer and evaluate three distinct models. Their deployment on VU13P FPGA chip achieved latency less than 2us, demonstrating the potential for real-time applications. HLS4ML compatibility with any TensorFlow-built transformer model further enhances the scalability and applicability of this work. Index Terms: FPGAs, machine learning, transformers, high energy physics, LIGO |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_05207 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Low Latency Transformer Inference on FPGAs for Physics Applications with hls4ml Jiang, Zhixing Yin, Dennis Chen, Yihui Khoda, Elham E Hauck, Scott Hsu, Shih-Chieh Govorkova, Ekaterina Harris, Philip Loncar, Vladimir Moreno, Eric A. Machine Learning This study presents an efficient implementation of transformer architectures in Field-Programmable Gate Arrays(FPGAs) using hls4ml. We demonstrate the strategy for implementing the multi-head attention, softmax, and normalization layer and evaluate three distinct models. Their deployment on VU13P FPGA chip achieved latency less than 2us, demonstrating the potential for real-time applications. HLS4ML compatibility with any TensorFlow-built transformer model further enhances the scalability and applicability of this work. Index Terms: FPGAs, machine learning, transformers, high energy physics, LIGO |
| title | Low Latency Transformer Inference on FPGAs for Physics Applications with hls4ml |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2409.05207 |