I hope you can create a command-line window similar to AudioSR.

[chencxt]

I hope you can create a command-line window similar to AudioSR.
From a regular player's perspective, we only interact with a dark window by “entering commands --> waiting for output.” Writing the library as a batch script results in too many possible outputs, making it impossible to batch-control generation quality. Therefore, I hope you, the author, can write a script for batch processing (using the arg library) so other users can simply enter commands to use it.
In fact, https://github.com/jarredou/FlashSR-Colab-Inference already accomplishes this, but it appears to have been inactive for over a year. I sincerely hope you can update this script. Currently, it still uses the original model weights instead of the compact 2MB model you're currently utilizing.

[ysharma3501]

@chencxt Yeah, I will try to do this in my next version. This is also a different FlashSR, it’s just named the same thing unfortunately. However, I will create command like version for my next version.