search

WindQAQ / listen-attend-and-spell

Tensorflow implementation of "Listen, Attend and Spell" authored by William Chan. This project utilizes input pipeline and estimator API of Tensorflow, which makes the training and evaluation truly end-to-end.
Apache License 2.0
89 stars 32 forks source link

Trouble in creating TfRecord file #8

Open ashu5644 opened 5 years ago

ashu5644 commented 5 years ago

Hi, I read LAS paper recently and found your implementation in tensorflow. I am trying to run your code but I am having problem in creating tfrecord file for TIMIT dataset. As I downloaded the dataset, it contains several files. I am not sure how to use them for creating tfrecord file and as I read (tfrecord reading and writing tutorial) before writing data into tfrecord file, it requires a dictionary format. so what type of format should I fed into tfrecord file to work with your code? And which files I should use for creating tfrecord for TIMIT dataset?

YashBangera7 commented 5 years ago

Were you successful in creating the Tfrecords file from TIMIT dataset? If so, please help me do so. It is very urgent and would help a great deal in my project. details of system: Python 3.6 tensorflow-gpu 1.8 nvidia gtx 1050ti gpu I have TIMIT dataset splitted into Test and Train dataset.