{"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/tensorquant-a-simulation-toolbox-for-deep","title":"TensorQuant - A Simulation Toolbox for Deep Neural Network Quantization","arxiv_id":"1710.05758","date":"2017-10-13","proceeding":null,"authors":["Dominik Marek Loroch","Norbert Wehn","Franz-Josef Pfreundt","Janis Keuper"],"abstract":"Recent research implies that training and inference of deep neural networks\n(DNN) can be computed with low precision numerical representations of the\ntraining/test data, weights and gradients without a general loss in accuracy.\nThe benefit of such compact representations is twofold: they allow a\nsignificant reduction of the communication bottleneck in distributed DNN\ntraining and faster neural network implementations on hardware accelerators\nlike FPGAs. Several quantization methods have been proposed to map the original\n32-bit floating point problem to low-bit representations. While most related\npublications validate the proposed approach on a single DNN topology, it\nappears to be evident, that the optimal choice of the quantization method and\nnumber of coding bits is topology dependent. To this end, there is no general\ntheory available, which would allow users to derive the optimal quantization\nduring the design of a DNN topology. In this paper, we present a quantization\ntool box for the TensorFlow framework. TensorQuant allows a transparent\nquantization simulation of existing DNN topologies during training and\ninference. TensorQuant supports generic quantization methods and allows\nexperimental evaluation of the impact of the quantization on single layers as\nwell as on the full topology. In a first series of experiments with\nTensorQuant, we show an analysis of fix-point quantizations of popular CNN\ntopologies.","url_abs":"http://arxiv.org/abs/1710.05758v1","url_pdf":"http://arxiv.org/pdf/1710.05758v1.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":"tensorquant-a-simulation-toolbox-for-deep","repo_url":"https://github.com/cc-hpc-itwm/TensorQuant","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"tf","reach":null},{"paper_slug":"tensorquant-a-simulation-toolbox-for-deep","repo_url":"https://github.com/DominikFHG/TensorQuant","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":null}],"tasks":[{"task_slug":"quantization","task_name":"Quantization"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1710.05758","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}