search

k2-fsa / k2

FSA/FST algorithms, differentiable, with PyTorch compatibility.
https://k2-fsa.github.io/k2
Apache License 2.0
1.11k stars 213 forks source link

Debug statements #1300

Closed Manjunath-mlp closed 2 weeks ago

Manjunath-mlp commented 2 weeks ago

Suppose if we made any changes in files (in src .cu,.h files),do i need to do the source installation again or is there any other way to reflect these changes .

csukuangfj commented 2 weeks ago

The first build is time-consuming, but later build is significantly faster. Please rebuild k2 from source if change any code.

You can follow the doc for how to build k2 from source.

By the way, if you don't change any CUDA-related code, then building a CPU-only version of k2 from source is super fast.

csukuangfj commented 2 weeks ago

Please don't download files from unknown sources.