Person Tube Retrieval via Language Description

Authors: Hehe Fan, Yi Yang10754-10761

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results on person tube retrieval via language description and other two related tasks demonstrate the efficacy of MSSP.
Researcher Affiliation Collaboration Hehe Fan,1,2 Yi Yang 2 1Baidu Research 2Re LER, CAI, University of Technology Sydney
Pseudocode No The paper does not contain any structured pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any concrete access information for open-source code.
Open Datasets Yes We evaluate variant MSSPs on the CUHK-PEDES datatset (Li et al. 2017b) and the MSR-VTT datatset (Xu et al. 2016). The training set has 34,054 images, 11,003 persons and 68,126 descriptions. We follow the same data split as 6,513, 2,990 and 497 clips in the training, testing and validation sets, respectively.
Dataset Splits Yes 5,500 persons are for training, 284 for validation and 284 for evaluation. The validation set has 3,078 images, 1,000 persons and 6,158 descriptions. We follow the same data split as 6,513, 2,990 and 497 clips in the training, testing and validation sets, respectively.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments.
Software Dependencies No The paper mentions We have implemented our algorithm using both Paddle Paddle and Py Torch, but does not specify version numbers for these software components.
Experiment Setup Yes Models are trained for 2,500 iterations, with batch size 1,500 and learning rate 0.01. Unless otherwise specified, λxy, λyx, λxx and λyy are set to 1.0, 2.0, 0.001 and 0.1, respectively.