Don't have Telegram yet? Try it now!
https://hgpu.org/?p=7568
Efficient Parallelization of Natural Language Applications using GPUs