We are looking for some developer experienced in the field of Text to speech. Development language should be python. Libraries which should be used are not fixed but we prefer tacotron2 in combination with waveglow. The engine should be able to work completely offline without any network connection. The quality should human realistic. Samples required of previous work.
The work should be done as Jupyter Notebooks to get better overview about the code and overview about the quality in asr. Your work will be splitted into different task.
-Load open data sets to server
-prepare script to globally rework the mapping, data preparation
-create good / strong data hierarchy for languages, types, phenoms
-Create Jupyter notebook for speech synthesizing
-create seperate jupyter notebook for training (the engine should be able to do trainings with checkpoints to later continue training with smaller amount of data without retraining ALL)
-create jupyter notebook for testing/demo
-create RESTful API with JSON output similiar to Google Speech Synthetic API (we mean structure, wording)
11 freelancere byder i gennemsnit $1334 på dette job
I can complete this within 14 days complete with integration to django rest framework for the restful api. Kindly contact me and we can discuss more and begin the project.