search

mozilla / DeepSpeech

DeepSpeech is an open source embedded (offline, on-device) speech-to-text engine which can run in real time on devices ranging from a Raspberry Pi 4 to high power GPU servers.
Mozilla Public License 2.0
25.44k stars 3.98k forks source link

Support for TensorFlow 2 #3706

Closed peterjungx closed 2 years ago

peterjungx commented 3 years ago

DeepSpeech not using TensorFlow 2 is a major limitation.

Today, Ubuntu 20.04 LTS is among the most used linux distributions for Deep Learning related development. Nvidia is at version 11.5 with its CUDA Toolkit, and does not provide version 10.2 of the toolkit for Ubuntu 20.04 LTS. Yet the TF version used by DeepSpeech requires CUDA Toolkit 10.2. This means DeepSpeech is not usable on one of the most popular linux distros, Ubuntu 20.04 LTS.

I could not find any roadmap of the DeepSpeech developers on this issue. I think clarifying the stance would be valuable here.

ftyers commented 3 years ago

Check out #3693.

peterjungx commented 3 years ago

@ftyers thank you! This explains a lot.