Closed vishaldhull09 closed 3 years ago
No, my demo works completely differently from theirs. Mine uses WASM instead of JS for RNNoise. To use your own model, change https://github.com/xiph/rnnoise/blob/master/src/rnn_data.c in the submodule, and then run https://github.com/wegylexy/rnnoise_wasm/blob/master/build-docker.sh to build a new WASM.
No, my demo works completely differently from theirs. Mine uses WASM instead of JS for RNNoise. To use your own model, change https://github.com/xiph/rnnoise/blob/master/src/rnn_data.c in the submodule, and then run https://github.com/wegylexy/rnnoise_wasm/blob/master/build-docker.sh to build a new WASM.
i was trying to use it directly in browser , I got this
Uncaught (in promise) TypeError: Failed to execute 'compile' on 'WebAssembly': Incorrect response MIME type. Expected 'application/wasm'.
do I have to change anything in code or build file ?
Your server must specify Content-Type: application/wasm
in the response header for *.wasm
files.
To test on your local machine, you can create a simple node.js
server.
See https://github.com/wegylexy/rnnoise_wasm/blob/master/demo/server.mjs#L25 .
For nginx
, config MIME types as follow:
types {
application/wasm wasm;
}
thanks for the help , i got this can we save and download input and denoised voice separately....or to upload a voice and get a denoised output
No downloading/uploading, it all runs in the browser locally. Do whatever you like.
https://jmvalin.ca/demo/rnnoise/noise.js
as far as i understood , demo for browser is using this file to load Rnnoise any idea how this file is created from the saved model, as I don't want to use node.js