search

Crystalsound / FRN

Other
26 stars 6 forks source link

Output for the Pre-trained onxx Model was 1KB #3

Closed ali7569392725 closed 3 weeks ago

ali7569392725 commented 1 year ago

Tried Inferencing the already existing model with 48KHz sample audio lossy file , the output length is expected to match with input length , but output was generating 1KB image

vietanh125 commented 1 year ago

Can you try this demo and see if it generates output correctly?

ali7569392725 commented 1 year ago

Demo is working fine but while I'm running the inference script , output is not properly generating ,Following is the command and modified my output & input paths in config.

"python inference_onnx.py --onnx_path lightning_logs/version_0/checkpoints/frn.onnx"

vietanh125 commented 10 months ago

Hi, sorry for the late reply.

Probably the onnx model in the repo is different from one in the HF space. Can you try using the one in HF space? https://huggingface.co/spaces/anhnv125/FRN/blob/main/lightning_logs/version_0/checkpoints/frn.onnx