Skip to content

Latest commit

 

History

History
29 lines (19 loc) · 1.03 KB

README.md

File metadata and controls

29 lines (19 loc) · 1.03 KB

TERC-ML - Team 4

Image Sets and Tags:

https://spark-xlab.slack.com/archives/C70DFLJ7M/p1510588534000126 https://spark-xlab.slack.com/archives/C70DFLJ7M/p1510588582000116 https://spark-xlab.slack.com/archives/C70DFLJ7M/p1511291170000494

Testing was done using the submit.sh bash script on scc.bu.edu. Place the raw images under ./data/raw and either use the script or run bottleneck_model.py. The weights from our model are located in the final results folder.

File descriptions:

bottleneck_model.py - where the model is datagenerator.py - code to transform images get_data.py - code for extracting and formating tags get_img_dict - file containing a function for mapping image names to their tags BU10000Set.csv - file containing tags for all images

Folder Descriptions:

logs - output from running the model on the SCC tests - collecton of runs using different parameters and model architectures final results - contains the output from our best run initial results - contains the output of our initial model deprecated - older, deprecated code