admin管理员组

文章数量:1399225

Trying to run the sample python code given here [(.py)] in Colab.

I tried runnning the same code in all cpu and gpus available in colab and I got the same error. The same code when run on my macbook intel on cpu gives no error. This code also surprisingly worked in kaggle.

code:

!pip install torch torchaudio -f .html --quiet
!pip install deepfilternet --quiet

from df.enhance import enhance, init_df, load_audio, save_audio
from df.utils import download_file

if __name__ == "__main__":
    # Load default model
    model, df_state, _ = init_df()
    # Download and open some audio file. You use your audio files here
    audio_path = download_file(
        ".wav",
        download_dir=".",
    )
    audio, _ = load_audio(audio_path, sr=df_state.sr())
    # Denoise the audio
    enhanced = enhance(model, df_state, audio)
    # Save for listening
    save_audio("enhanced.wav", enhanced, df_state.sr())

complete error:

2025-03-25 18:11:11 | INFO     | DF | Loading model settings of DeepFilterNet3
2025-03-25 18:11:11 | INFO     | DF | Using DeepFilterNet3 model at /root/.cache/DeepFilterNet/DeepFilterNet3
2025-03-25 18:11:11 | INFO     | DF | Initializing model `deepfilternet3`
2025-03-25 18:11:11 | INFO     | DF | Found checkpoint /root/.cache/DeepFilterNet/DeepFilterNet3/checkpoints/model_120.ckpt.best with epoch 120
2025-03-25 18:11:11 | INFO     | DF | Running on device cuda:0
2025-03-25 18:11:11 | INFO     | DF | Model loaded
---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
[<ipython-input-13-8e53b6efe558>](https://localhost:8080/#) in <cell line: 0>()
     12     audio, _ = load_audio(audio_path, sr=df_state.sr())
     13     # Denoise the audio
---> 14     enhanced = enhance(model, df_state, audio)
     15     # Save for listening
     16     save_audio("enhanced.wav", enhanced, df_state.sr())

2 frames
[/usr/local/lib/python3.11/dist-packages/torch/utils/_contextlib.py](https://localhost:8080/#) in decorate_context(*args, **kwargs)
    114     def decorate_context(*args, **kwargs):
    115         with ctx_factory():
--> 116             return func(*args, **kwargs)
    117 
    118     return decorate_context

[/usr/local/lib/python3.11/dist-packages/df/enhance.py](https://localhost:8080/#) in enhance(model, df_state, audio, pad, atten_lim_db)
    232         audio = F.pad(audio, (0, n_fft))
    233     nb_df = getattr(model, "nb_df", getattr(model, "df_bins", ModelParams().nb_df))
--> 234     spec, erb_feat, spec_feat = df_features(audio, df_state, nb_df, device=get_device())
    235     enhanced = model(spec.clone(), erb_feat, spec_feat)[0].cpu()
    236     enhanced = as_complex(enhanced.squeeze(1))

[/usr/local/lib/python3.11/dist-packages/df/enhance.py](https://localhost:8080/#) in df_features(audio, df, nb_df, device)
    188 
    189 def df_features(audio: Tensor, df: DF, nb_df: int, device=None) -> Tuple[Tensor, Tensor, Tensor]:
--> 190     spec = df.analysis(audio.numpy())  # [C, Tf] -> [C, Tf, F]
    191     a = get_norm_alpha(False)
    192     erb_fb = df.erb_widths()

TypeError: argument 'input': 'ndarray' object cannot be converted to 'PyArray<T, D>'

Any help is appreciated!

本文标签：

版权声明：本文标题：audio - Deepfilter example python code in Colab gives error - TypeError: argument &#39;input&#39;: &#39;ndarray& 内容由网友自发贡献，该文观点仅代表作者本人， 转载请联系作者并注明出处：http://www.betaflare.com/web/1744175815a2593993.html， 本站仅提供信息存储空间服务，不拥有所有权，不承担相关法律责任。如发现本站有涉嫌抄袭侵权/违法违规的内容，一经查实，本站将立刻删除。