Skip to content

Latest commit

 

History

History
34 lines (19 loc) · 750 Bytes

dawnbench.md

File metadata and controls

34 lines (19 loc) · 750 Bytes
Raw

Dawnbench instructions

  1. Use Amazon Deep Learning AMI (Ubuntu) Version 15.0 - ami-0b43cec40e1390f34

  2. Configure Drive for 100GB of space + SSD

  3. Install pytorch 0.4.1 w/ CUDA:

conda install pytorch torchvision cuda92 -c pytorch

Clone git repo:

git clone https://github.com/hitvoice/DrQA.git

Follow readme:

cd DrQA

pip install -r requirements.txt

bash download.sh

python prepro.py

python train.py -e 40 -bs 128 --save_last_only --save_dawn_logs

Technically, the p3 can support larger batch sizes, but convergence decreases, so use this size.

Then, use the following to generate the .tsv entry for Dawnbench:

grep dawn_entry models/log.txt | cut -c 33- > models/dawn_entry.tsv