Pre-train model release about synthetic-voice-detection-vocoder-artifacts

[csun22]

Thank you all. The weight of the trained model has been uploaded to the readme with a Google Drive link.
https://drive.google.com/file/d/1TWdsCFKP2luAfhpB91N9X4z1gsJMvvhI/view?usp=drive_link
Please let me know if you have any questions.

[csun22]

Hi Steven，
Do you mean our pre-train model?

[steven84418]

The model you used for the result

[csun22]

Yes. There are two ways you can do it.
1.python eval.py --input_path /your/path/to/sample.wav --model_path /your/path/to/your_model.pth the eval.py file is provided.
2. https://zinc.cse.buffalo.edu/ubmdfl/deep-o-meter/landing_page
using this platform to test out.

[Rizwanali324]

can u upload weights here instead of on website as open source

[jreivilo]

Hey

Would be great indeed to have the weight of the trained model in the repo :)

Thanks

[Rizwanali324]

Thank you all. The weight of the trained model has been uploaded to the readme with a Google Drive link. https://drive.google.com/file/d/1TWdsCFKP2luAfhpB91N9X4z1gsJMvvhI/view?usp=drive_link Please let me know if you have any questions.

any instruction for to use it for custom as I'm working on project with this dataset

[Rizwanali324]

Also, can you share other evaluation matrices?
we are trying to deploy it on real time so we need to know inference speed and percision

[SwenHou]

Thank you all. The weight of the trained model has been uploaded to the readme with a Google Drive link. https://drive.google.com/file/d/1TWdsCFKP2luAfhpB91N9X4z1gsJMvvhI/view?usp=drive_link Please let me know if you have any questions.

any instruction for to use it for custom as I'm working on project with this dataset

To use the pretrained weight, you can run 'python eval.py --input_path /your/path/to/sample.wav --model_path /your/path/to/your_model.pth'

you can also check the code in eval.py to make any modifications according to your custom requirements

[SwenHou]

Also, can you share other evaluation matrices? we are trying to deploy it on real time so we need to know inference speed and percision

You can check our paper for any evaluation metrices, the inference speed is fast enough for real time detection, inference time should be less than 1s for about 10s of speech, after loading the weight.

[Rizwanali324]

Also, can you share other evaluation matrices? we are trying to deploy it on real time so we need to know inference speed and percision

You can check our paper for any evaluation metrices, the inference speed is fast enough for real time detection, inference time should be less than 1s for about 10s of speech, after loading the weight.

there's is no such matrix just reduce error rate