Skip to content

Latest commit

 

History

History
25 lines (13 loc) · 977 Bytes

README.md

File metadata and controls

25 lines (13 loc) · 977 Bytes
Raw

quickdraw-chainer

Google Quick, Draw! doodle recognition model implementation in Chainer

Open source for 11th Software Innovation Contest.

Files

  • download.sh, filelist.txt: fetch simplifed datasets from Google Cloud Platform

  • dataset.py: Preprocess the datasets (selection and division)

  • model.py: Model definition, supports CPU and GPU (switch by gpu_id)

  • train.py: Main functions for data reading and training

  • PadIterator.py: A special version of SerialIterator for padding the batch and pass its length into model

  • /demo: see /demo/README.md

Usage

Execute download.sh to get dataset (you may need to modify it to store them in a proper folder), then dataset.py to preprocess, train.py to start training.

Logs, visualization and snapshots will automatically save in /result.