GitHub - backpropper/ContextAwareSpeechRecognition

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
Caffe and rest		Caffe and rest
object		object
place		place
.DS_Store		.DS_Store
README		README
Writeup		Writeup
av.pdf		av.pdf
newrnn.py		newrnn.py

Repository files navigation

#ContextAwareSpeechRecognition

Details about the project can be found in Writeup

To run a single fold (including video features)---here 1 represents first fold can be (1-5):
qsub -q gpu run_video1.sh      

To run a single fold (baseline)---here 1 represents first fold can be (1-5):
qsub -q gpu run_baseline1.sh

Place Features present in /data/ASR5/abhinav5/PlacesAlexNet_360h
Object Features present in /data/ASR5/abhinav5/YTubeV2_360h
Data present in /data/ASR5/abhinav5/train_360h