{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/deepdsl-a-compilation-based-domain-specific","title":"DeepDSL: A Compilation-based Domain-Specific Language for Deep Learning","arxiv_id":"1701.02284","date":"2017-01-09","proceeding":null,"authors":["Tian Zhao","Xiaobing Huang","Yu Cao"],"abstract":"In recent years, Deep Learning (DL) has found great success in domains such\nas multimedia understanding. However, the complex nature of multimedia data\nmakes it difficult to develop DL-based software. The state-of-the art tools,\nsuch as Caffe, TensorFlow, Torch7, and CNTK, while are successful in their\napplicable domains, are programming libraries with fixed user interface,\ninternal representation, and execution environment. This makes it difficult to\nimplement portable and customized DL applications.\n  In this paper, we present DeepDSL, a domain specific language (DSL) embedded\nin Scala, that compiles deep networks written in DeepDSL to Java source code.\nDeep DSL provides (1) intuitive constructs to support compact encoding of deep\nnetworks; (2) symbolic gradient derivation of the networks; (3) static analysis\nfor memory consumption and error detection; and (4) DSL-level optimization to\nimprove memory and runtime efficiency.\n  DeepDSL programs are compiled into compact, efficient, customizable, and\nportable Java source code, which operates the CUDA and CUDNN interfaces running\non Nvidia GPU via a Java Native Interface (JNI) library. We evaluated DeepDSL\nwith a number of popular DL networks. Our experiments show that the compiled\nprograms have very competitive runtime performance and memory efficiency\ncompared to the existing libraries.","url_abs":"http://arxiv.org/abs/1701.02284v1","url_pdf":"http://arxiv.org/pdf/1701.02284v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"deepdsl-a-compilation-based-domain-specific","repo_url":"https://github.com/deepdsl/deepdsl","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"tf","reach":null}],"tasks":[{"task_slug":"deep-learning","task_name":"Deep Learning"},{"task_slug":null,"task_name":"GPU"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}